Neuromodulation: Technology at the Neural Interface

Received: March 4, 2016 Revised: May 23, 2016 Accepted: May 23, 2016

(onlinelibrary.wiley.com) DOI: 10.1111/ner.12538

The Polyanalgesic Consensus Conference

(PACC): Recommendations on Intrathecal Drug
Infusion Systems Best Practices and Guidelines
Timothy R. Deer, MD*; Jason E. Pope, MD†; Salim M. Hayek, MD, PhD‡;
Anjum Bux, MD§; Eric Buchser, MD¶; Sam Eldabe, MD**;
Jose A. De Andre s, MD, PhD, EDRA††; Michael Erdek, MD‡‡;
Dennis Patin, MD§§; Jay S. Grider, PhD, MBA¶¶; Daniel M. Doleys, PhD***;
Marilyn S. Jacobs, PhD†††; Tony L. Yaksh, PhD‡‡‡;
Lawrence Poree, MD, PhD§§§; Mark S. Wallace, MD¶¶¶; Joshua Prager, MD****;
Richard Rauck, MD††††; Oscar DeLeon, MD‡‡‡‡; Sudhir Diwan, MD§§§§;
Steven M. Falowski, MD¶¶¶¶; Helena M. Gazelka, MD*****;
Philip Kim, MD†††††‡‡‡‡‡; Michael Leong, MD§§§§§;
Robert M. Levy, MD, PhD¶¶¶¶¶; Gladstone McDowell II, MD******;
Porter McRoberts, MD††††††; Ramana Naidu, MD‡‡‡‡‡‡;
Samir Narouze, MD, PhD§§§§§§; Christophe Perruchoud, MD¶¶¶¶¶¶;
Steven M. Rosen, MD*******; William S. Rosenberg, MD†††††††;
Michael Saulino, MD, PhD‡‡‡‡‡‡‡; Peter Staats, MD§§§§§§§¶¶¶¶¶¶¶;
Lisa J. Stearns, MD********; Dean Willis, MD††††††††; Elliot Krames, MD‡‡‡‡‡‡‡‡;
Marc Huntoon, MD§§§§§§§§; Nagy Mekhail, MD, PhD¶¶¶¶¶¶¶¶
Introduction: Pain treatment is best performed when a patient-centric, safety-based philosophy is used to determine an algorithmic process to
guide care. Since 2007, the International Neuromodulation Society has organized a group of experts to evaluate evidence and create a Polyanalge-
sic Consensus Conference (PACC) to guide practice.
Methods: The current PACC update was designed to address the deficiencies and innovations emerging since the previous PACC publication of
2012. An extensive literature search identified publications between January 15, 2007 and November 22, 2015 and authors contributed additional
relevant sources. After reviewing the literature, the panel convened to determine evidence levels and degrees of recommendations for intrathecal
therapy. This meeting served as the basis for consensus development, which was ranked as strong, moderate or weak. Algorithms were developed
for intrathecal medication choices to treat nociceptive and neuropathic pain for patients with cancer, terminal illness, and noncancer pain, with
either localized or diffuse pain.
Results: The PACC has developed an algorithmic process for several aspects of intrathecal drug delivery to promote safe and efficacious
evidence-based care. Consensus opinion, based on expertise, was used to fill gaps in evidence. Thirty-one consensus points emerged from the
panel considerations.
Conclusion: New algorithms and guidance have been established to improve care with the use of intrathecal drug delivery.

Address correspondence to: Timothy R. Deer, MD, President and CEO, The Center for Pain Relief, Inc., 400 Court Street, Ste 100, Charleston, WV, USA. Email: DocTDeer@
aol.com

* Center for Pain Relief, Charleston, WV, USA;

†
Summit Pain Alliance, Santa Rosa, CA, USA;
‡
Case Western Reserve University, Cleveland, OH, USA;
§
Anesthesia and Chronic Pain Management, Ephraim McDowell Regional Medical Center, Danville, KY, USA;
¶
Anaesthesia and Pain Management Department, EHC Hosptial, Morges, and CHUV University Hospital, Lausanne, Switzerland;
** The James Cook University Hospital, Middlesbrough, UK;
††
Valencia School of Medicine, Hospital General Universitario, Valencia, Spain;
‡‡
Anesthesiology and Critical Care Medicine, Johns Hopkins University School of Medicine, Baltimore, MD, USA;
§§
University of Miami Health System, Miami, FL, USA;
¶¶
University of Kentucky College of Medicine, UK HealthCare Pain Services, Lexington, KY, USA;
*** Doleys Clinic, Birmingham, AL, USA;
†††
California Pain Medicine Center, Los Angeles, CA, USA;
96

‡‡‡
Anesthesiology and Pharmacology, University of California, San Diego, CA, USA;

www.neuromodulationjournal.com C 2017 International Neuromodulation Society

V Neuromodulation 2017; 20: 96–132
 INTRATHECAL THERAPY BEST PRACTICES AND GUIDELINES

Keywords: Chronic pain, consensus, fixed rate pump, intrathecal drug delivery, neuropathic pain, nonmalignant pain, opioid,
programmable pump, psychological evaluation, safety
Conflict of Interest: Dr. Deer is a consultant for St. Jude Medical Inc., Medtronic Inc., Bioness Inc., Vertos Medical Inc., Nevro Corp., Flowonix Medical Inc.,
Axonics Ltd., Ethos, Spinal Therapeutics, Saluda Medical Pty. Ltd., and Nuvectra Corp. He serves on the Advisory Board of St. Jude Medical Inc., Medtronic
Inc., Bioness Inc., Nevro Corp., Flowonix Medical Inc., Jazz Pharmaceuticals PLC, and Axonics Ltd. He is also a speaker for Jazz Pharmaceuticals PLC, and has
stock options in Bioness Inc., Vertos Medical Inc., Axonics Ltd., Spinal Therapeutics, Saluda Medical Pty. Ltd., and Nuvectra Corp. Dr. Pope is a consultant for
Medtronic Inc., St. Jude Medical Inc., Jazz Pharmaceuticals PLC, Nevro Corp., Nuvectra Corp., and Flowonix Medical Inc. Dr. Hayek is a consultant for Flowo-
nix Medical Inc. and Mallinckrodt Pharmaceuticals. Dr. Bux receives speaking honoraria from Medtronic Inc., and Jazz Pharmaceuticals PLC. Dr. Buchser is a
consultant and researcher for Medtronic Inc. Dr. Eldabe is a consultant for Medtronic Inc. Dr. De Andr es has no conflicts of interest to report. Dr. Erdek has
no conflicts of interest to report. Dr. Patin is a consultant for Medtronic Inc. Dr. Grider is a consultant for Interlink Spine, and is a speaker for Medtronic Inc.
Dr. Doleys is a speaker for Medtronic Inc. and Kal eo Pharma/Evzio. Dr. Jacobs has no conflicts of interest to report. Dr. Yaksh is a consultant for Adynxx, and
a researcher for Medtronic Inc., Flowonix Medical Inc., and Jazz Pharmaceuticals PLC. Dr. Poree is on the Medtronic Inc. Safety Advisory Panel. He is a consul-
tant for Medtronic Inc., Nalu Medical Inc., Stimwave Technology Ltd., St. Jude Medical Inc., and Mallinckrodt Pharmaceuticals. Dr. Wallace is a consultant for
Jazz Pharmaceuticals PLC and the Alfred Mann Foundation. Dr. Prager is a speaker for Insys Therapeutics and a consultant for Medtronic Inc. Dr. Rauck
advises and conducts research for Jazz Pharmaceuticals PLC and Medtronic Inc. He is also a speaker for Jazz Pharmaceuticals PLC. Dr. De Leon-Casasola is a
consultant for Mallinckrodt Pharmaceuticals, Purdue Pharma LP, and Depomed Inc. Dr. Diwan has no conflicts of interest to report. Dr. Falowski is a consul-
tant, researcher, and speaker for Medtronic Inc. and St. Jude Medical Inc. He is a consultant and speaker for Nevro Corp., and a researcher for Saluda
Medical Pty. Ltd. Dr. Gazelka has no conflicts of interest to report. Dr. Kim is a consultant and speaker for Medtronic Inc. and Jazz Pharmaceuticals PLC, and
is a consultant for Biotronic NeuroNetwork. Dr. Leong is a consultant for Boston Scientific Corp., Jazz Pharmaceuticals PLC, and Sorrento Therapeutics. Dr.
Levy is a minority shareholder in Spinal Modulation Inc., Bioness Inc., Vertos Medical Inc., and Nevro Corp. He is a consultant for Medtronic Inc., St. Jude
Medical Inc., Spinal Modulation Inc., Nevro Corp., Saluda Medical Pty. Ltd., and Flowonix Medical Inc. Dr. McDowell is an advisor and consultant for Med-
tronic Inc., Jazz Pharmaceuticals PLC, and Flowonix Medical Inc. Dr. McRoberts is an advisor for St. Jude Medical Inc., Medtronic Inc., Flowonix Medical Inc.,
Nalu Medical Inc., Vertiflex Inc., Boston Scientific Corp., Bioness Inc, SPR, Interlink Spine and OrthoSensor. He is a researcher for Mesoblast, Medtronic Inc.,
Vertiflex Inc., and Mainstay Medical. He is a speaker for Nevro Corp., and St. Jude Medical Inc. Dr. Naidu is a consultant and speaker for Halyard Health. Dr.
Narouze is a consultant for St. Jude Medical Inc. Dr. Perruchoud is a consultant, researcher, and speaker for Medtronic Inc. He is a consultant and researcher
for Spinal Modulation Inc., and is a consultant for St. Jude Medical Inc. Dr. Rosen is a consultant and researcher for Flowonix Medical Inc. and Saluda Medical
Pty. Ltd., and a researcher for Nevro Corp. Dr. Rosenberg is a consultant and speaker for Medtronic Inc. and St. Jude Medical Inc. Dr. Saulino is a researcher
and speaker for Medtronic Inc., Mallinckrodt Pharmaceuticals and Jazz Pharmaceuticals PLC. Dr. Staats is a consultant for ElectroCore Medical LLC, St. Jude
Medical Inc., Boston Scientific Corp., and Medtronic Inc. He does research for Bioness Inc., Spinal Modulation Inc., Boston Scientific Corp., Vertos Medical
Inc., St. Jude Medical Inc., Saluda Medical Pty. Ltd., Gr€
uenthal GmbH, and ElectroCore Medical LLC. He receives product royalty and licensing fees from Neu-
rogesX Inc., and has relevant holdings in ElectroCore Medical LLC. Dr. Stearns is a consultant for Medtronic Inc., Spinal Modulation Inc., Pfizer Inc., and the
Alfred Mann Foundation. Dr. Willis has no conflicts to disclose. Dr. Krames has no conflicts to report. Dr. Huntoon is a researcher for CNS Therapeutics and a
part-time employee of Regional Anesthesia and Pain Medicine. Dr. Mekhail is an advisor for Flowonix Medical Inc., Boston Scientific Corp. and Saluda Medical
Pty. Ltd. He is a researcher for Mesoblast Ltd., Halyard Health, CNS Therapeutics, Axsome Therapeutics and Stimwave Technology Ltd.

§§§
Pain Clinic of Monterey Bay, University of California at San Francisco, San Francisco, CA, USA;
¶¶¶
Department of Anesthesia, UC San Diego, San Diego, CA, USA;
**** Center for the Rehabilitation Pain Syndromes (CRPS) at UCLA Medical Plaza, Los Angeles, CA, USA;
††††
Carolina Pain Institute, Wake Forest Baptist Health, Winston-Salem, NC, USA;
‡‡‡‡
Roswell Park Cancer Institute, SUNY, Buffalo, NY, USA;
§§§§
Manhattan Spine and Pain Medicine, Lenox Hill Hospital, New York, NY, USA;
¶¶¶¶
St. Luke’s University Health Network, Bethlehem, PA, USA;
***** Mayo Clinic, Rochester, MN, USA;
†††††
Bryn Mawr Hospital, Bryn Mawr, PA, USA;
‡‡‡‡‡
Christiana Hospital, Newark, DE, USA;
§§§§§
Stanford University, Palo Alto, CA, USA;
¶¶¶¶¶
Marcus Neuroscience Institute, Boca Raton, FL, USA;
****** Integrated Pain Solutions, Columbus, OH, USA;
††††††
Holy Cross Hospital, Fort Lauderdale, Florida, USA;
‡‡‡‡‡‡
San Francisco Medical Center, University of California, San Francisco, CA, USA;
§§§§§§
Summa Western Reserve Hospital, Cuyahoga Falls, OH, USA;
¶¶¶¶¶¶
H^opital de Zone Morges, Morges, Switzerland;
******* Fox Chase Pain Management Associates PC, Doylestown, PA, USA;
†††††††
Center for Relief of Pain, Kansas City, MO, USA;
‡‡‡‡‡‡‡
MossRehab, Elkins Park, PA, USA;
§§§§§§§
Premier Pain Management Centers, Shrewsbury, NJ, USA;
¶¶¶¶¶¶¶
Johns Hopkins University, Baltimore, MD, USA;
******** Center for Pain and Supportive Care, Phoenix, AZ, USA;
††††††††
Alabama Pain Center, Huntsville, AL, USA;
‡‡‡‡‡‡‡‡
Pacific Pain Treatment Center (ret.), San Francisco, CA, USA;
§§§§§§§§
Vanderbilt University Medical Center, Nashville, TN, USA; and
¶¶¶¶¶¶¶¶
Cleveland Clinic, Cleveland, OH, USA

For more information on author guidelines, an explanation of our peer review process, and conflict of interest informed consent policies, please go to http://www.wiley.
com/WileyCDA/Section/id-301854.html
97

[Correction added on 19 January 2017, after first online publication: the name of the third author has been corrected from “Salim Hayek” to “Salim M. Hayek”.]

www.neuromodulationjournal.com C 2017 International Neuromodulation Society

V Neuromodulation 2017; 20: 96–132
 DEER ET AL.

INTRODUCTION AND RATIONALE tailor algorithmic approaches to patient-specific characteristics that

The use of intraspinal (intrathecal [IT]) infusion of analgesic medica-
tions to treat patients with chronic refractory pain has increased since
its inception in the 1980s, and the need for clinical and outcomes
research in IT therapy is ongoing. New IT devices have been recently
introduced, along with novel chronic infusion strategies (1). Thus far,
research has not kept pace with the growing need for innovative IT
pain management, and clinical care and decision making have largely
relied on best evidence and expert opinion (2). Therefore, a consen-
sus opinion is needed to identify the current research and address
deficiencies in the data. Furthermore, as new IT therapies become
available, the need to refine patient selection and pain care algo-
rithms is required (3). With more than 80% of IT therapy in the United
States employed as off-label (4,5), there is a continuing need to help
navigate careful decision-making surrounding IT therapy.
Consensus Point 1. An update of the best practices of IT drug
delivery is needed due to many changes in patient care since the
last version of this living document.
Consensus Point 2. Evidence for practice guidance has been
improved since 2012, but additional expert guidance is needed to
fill the current gaps in clinical practice.
The Polyanalgesic Consensus Conference (PACC) panel was
formed to improve the safety and efficacy of IT therapy. PACC was
initiated in 2000 to address the research gaps and review the existing
data regarding IT therapy (2). This expert panel was composed of
health care providers within the field of IT therapy. The panel devel-
oped an IT drug selection algorithm on the basis of best evidence
and expert opinion, prepared supplemental reports that included a
relevant literature review, reported on results from a survey of peers
engaged in IT therapy, and described future directions in the field of
IT therapy. The PACC panel reconvened in 2003 (6) and 2007 (7) to
evaluate the most up-to-date literature and to update the algorithm
for intraspinal drug selection. In 2011, to formulate consensus opin-
ions on critical issues involving IT therapy and identify areas for
future research in the field, the PACC panel again convened, adding
a supplement on IT granuloma and describing new insights on rec-
ommended maximal concentrations and daily doses of IT agents (8).
Since then, renewed interest regarding noncancer pain management
employing IT agents has resurfaced, along with new interests on
infusion schema. This present update focuses on redefining patient-
centric trialing strategies and the recommended observation period
for trialing, algorithmic care for the near end-of-life patient, and the
importance of catheter position congruent with regional location of
the pain. The overarching theme of this present PACC update is to
continue to modify and adapt this living document.
Brief History of PACC
In spite of thorough discussions of various factors impacting
likely impact IT therapy (2,6–9). A number of clinical factors play
important roles in shaping specific IT interventions and medication
choices. These factors have been previously described (10), and
include patient diagnoses and expected survival time (11), patient
age (12,13), previous exposure to opioids (13–15), location of pain
(diffuse vs. localized), type of pain (nociceptive, neuropathic, or
mixed), the physiochemical properties of lipid solubility of the IT
medications employed (16,17), cerebrospinal fluid (CSF) flow
dynamics and pharmacokinetics (18,19), IT catheter location (20),
pump and catheter characteristics, kinetics of the IT infusate (20),
and psychological status (21–23) of the patient with chronic pain. A
detailed discussion of these factors will be entertained in the follow-
ing sections. Each component to be considered requires careful and
deliberate attention. Interestingly, unpublished industry reports
suggest that many practicing physicians in the United States and
Europe do not follow the PACC guidance of 2012 (24). Physicians
and scientists have been impacted by the advice given in this
process, with nearly 100 citations of the 2012 guidance. The effort
of the 2016 PACC is to provide recommendations based on
evidence and consensus and to continue to disseminate best
practice insights to practicing physicians worldwide.
Consensus Point 3. The 2016 PACC will continue the historical
goal of improving safety and efficacy of the global use of IT therapies.
METHODS
The PACC of 2016 was designed to address the deficiencies and
innovations emerging since the previous PACC of 2012 regarding IT
therapy. Participants were chosen based on an executive panel from
the International Neuromodulation Society (INS), with participants
from previous PACC guidelines automatically nominated. Other
nominations were made by board members based on a needs
assessment of topics to be addressed. All participants were identi-
fied to have an area of needed expertise, which could include exten-
sive experience in IT drug device management, basic science
research, clinical studies, or expertise in evidence assessment or pub-
lication. Invitations were subsequently sent to potential participants
and accepted prior to formal engagement. Meetings were held peri-
odically during the composition and drafting of the manuscript,
with meetings to rank evidence and develop consensus surrounding
IT therapy, as defined below. The authorship publication standards
outlined by the journal Neuromodulation and Wiley Publishing gov-
erned the working consensus group.
Literature Search Methods
A broad literature search was conducted to identify preclinical
and clinical data on IT therapy published from January 15, 2007,
through November 22, 2015. MEDLINEV R , BioMed CentralV
R , Current

patient responses to IT therapy, previous PACC versions did not Contents ConnectV R, EmbaseTM, International Pharmaceutical

Table 1. Hierarchy of Studies by the Type of Design (U.S. Preventive Services Task Force, Ref [25]).

Evidence level Study type

I At least one controlled and randomized clinical trial, properly designed

II-1 Well-designed, controlled, nonrandomized clinical trials
II-2 Cohort or case studies and well-designed controls, preferably multicenter
II-3 Multiple series compared over time, with or without intervention, and surprising results in noncontrolled experiences
III Clinical experience-based opinions, descriptive studies, clinical observations or reports of expert committees.
98

www.neuromodulationjournal.com C 2017 International Neuromodulation Society

V Neuromodulation 2017; 20: 96–132
 INTRATHECAL THERAPY BEST PRACTICES AND GUIDELINES

Table 2. Meaning of Recommendation Degrees (U.S. Preventive Services Task Force, Ref [25]).

Degree of recommendation Meaning

A Extremely recommendable (good evidence that the measure is effective and benefits outweigh the harms)
B Recommendable (at least, moderate evidence that the measure is effective and benefits exceed harms)
C Neither recommendable nor inadvisable (at least moderate evidence that the measure is effective, but benefits are
similar to harms and a general recommendation cannot be justified)
D Inadvisable (at least moderate evidence that the measure is ineffective or that the harms exceed the benefits)
I Insufficient, low quality or contradictory evidence; the balance between benefit and harms cannot be determined.

AbstractsVR , and Web of ScienceVR , Google Scholar, and Pubmed data
bases were searched for publications on a range of medications that
are either currently in use or potentially useful for the IT treatment
of chronic pain. Search terms included “intrathecal, intraspinal, mor-
phine, fentanyl, sufentanil, methadone, adenosine, hydromorphone,
meperidine, gabapentin, baclofen, ketorolac, midazolam, neostig-
mine, octreotide, ziconotide, ropivacaine, dexmedetomidine, cloni-
dine, bupivacaine, and lidocaine.” Each author performed
independent literature searches and the information was cross-
referenced and compiled for evidence analysis and consensus
review. These searches yielded 391 articles, which were examined
for relevance to the IT treatment of chronic pain. Google Scholar
was again searched for recent relevant information regarding IT
therapy for chronic pain, and additional literature considered by
panel members to be relevant to this new consensus paper was
reviewed. Wherever pertinent, proposed mechanisms of action for
the particular drug class are provided, along with a summary of pre-
clinical studies, followed by critical review. Literature published
before the dates stated above is cited when relevant. After review-
ing the literature, the PACC panel convened to develop

99

Figure 1. Contributor evidence assessment.

www.neuromodulationjournal.com C 2017 International Neuromodulation Society

V Neuromodulation 2017; 20: 96–132
 DEER ET AL.

opinions and recommendations offered are not intended to pro-

Table 3. Strength of Consensus.
Strength of consensus Definition*
Strong >80% consensus
Moderate 50–79% consensus
Weak <49% consensus
*Quorum defined as 80% of participants available for vote.
recommendations for IT analgesia. Supporting literature is included
following these recommendations and discussions of the panel.
Evidence-Based Analysis vs. Opinion
Similar to the Neuromodulation Appropriateness Consensus Con-
ference (NACC) of the International Neuromodulation Society publi-
cation in 2014, and fostered from the previous PACC statements, the
goal of this present PACC effort was to create a living document,
with continued refreshment and evidence synthesis ongoing, as
appropriate. Unlike the PACC of 2012, the effort of the 2016 PACC
was to apply a validated evidence-ranking system, outlining man-
agement and implementation of IT therapy. This effort was under-
scored by unpublished survey data suggesting poor adoption of the
previous proposed algorithms (personal communication with Med-
tronic plc. and Jazz Pharmaceuticals). It is clear, however, that IT
granuloma identification and management were significantly
impacted and improved on since the PACC of 2011 supplement was
published (24).
mote off-label uses of medications and devices. Additionally, these
recommendations should not be construed as a standard of care.
We will explore an evidence-based algorithm of pain care, patient
selection, drug selection, trialing strategies, implantation, and con-
centration and dosing. Physicians should consult their national
approval processes when making clinical decisions.
It is important to address the conflicting nature of evidence and
the need for consensus. Evidence and consensus are not mutually
exclusive, which may be the perception at first glance. Evidence
assessment, regardless of the strength, needs interpretation for clini-
cal application whenever used.
RECOMMENDATIONS OF PACC 2016
In this manuscript, we will explore the evidence-weighted and
consensus recommendations of the PACC regarding the following
topics:
• Evidence assessment
• Pain care algorithms
• Disease-specific indications and considerations
• Patient-selection considerations
• Medication-selection recommendations and considerations
• Use context of neuropathic and nociceptive pain
• Recommended starting dosages
• Variables affecting chronic intrathecal therapy
• Conclusions

Evidence Ranking
The United States Preventative Services Task Force (USPSTF) creat-
ed hierarchies of studies and degrees of recommendations based on
evidence rankings as outlined in Tables 1 and 2 (25). The PACC of
2016 has adopted these classifications, just as the NACC previously
adopted weighted recommendations for neurostimulation.
Authors of this manuscript were asked to complete reference
forms for their section’s assessment (Fig. 1). These forms were
then reviewed by the executive committee of the working group
and averaged. They served as the basis for review and consensus
development. The working group developed recommendations
based on evidence ranking, or consensus when evidence was
lacking, followed by assigning consensus rankings. The consensus
determination was performed during in-person meetings or via
teleconference with a quorum of 80% of the contributing authors
determining recommendation strength. Consensus rankings were
outlined as strong, moderate, or weak based on agreement, as
defined in Table 3.
As with any guideline, this document serves as a recommendation
regarding implementation and management of IT therapy. The
EVIDENCE ASSESSMENT
It is generally regarded that IT therapy offers a reliable, accurate,
safe, and efficacious treatment for both cancer and noncancer pain,
as well as for end-of-life pain care. There have been multiple reviews
discussing the efficacy and safety of IT therapy (26–29). Recently, IT
therapy options came under scrutiny by the state of California, with
success of continued access available only after a demonstration of
evidence was ruled favorable (30). A more thorough understanding
of the pharmacokinetic properties of IT medications (20) and CSF
flow dynamics within the IT space was the catalyst in creating better
IT therapy strategies (31). Furthermore, as defined previously, there
is evidence to support its use (26), based on the USPSTF criteria for
data ranking and evidence strength. USPSTF strength of evidence
for IT therapy was level II-3 for noncancer pain and level II-2 for can-
cer pain. A best practice article was written in 2014 by Prager et al.,
providing evidence in support of IT therapy and the needed place-
ment of IT therapy in the algorithm for cancer and noncancer pain
(27) (Table 4). The support for efficacy of IT opioid administration for
the management of chronic noncancer pain comes largely from pro-
spective and retrospective noncontrolled trials. Ziconotide has been

Table 4. Recommendations for Evidence Assessment of Intrathecal Therapy by the PACC Using USPSTF Criteria.

Statement Evidence level Recommendation grade Consensus level

Intrathecal therapy should be utilized for active cancer-related pain with opioids. I A Strong
Intrathecal therapy should be utilized for active cancer-related pain with ziconotide. I A Strong
Intrathecal therapy should be utilized for noncancer-related pain with opioids. III B Strong
100

Intrathecal therapy should be utilized for noncancer-related pain with ziconotide. I A Strong

www.neuromodulationjournal.com C 2017 International Neuromodulation Society

V Neuromodulation 2017; 20: 96–132

Figure 2. Algorithm for placement within the pain care algorithm for noncancer or non-end-of-life pain. DRG, dorsal root ganglion; HF10, high frequency stimula-
tion; PNfS, peripheral nerve field stimulation; PNS, peripheral nerve stimulation; SCS, spinal cord stimulation. Green arrows indicate affirmation or positive response;
red arrows signify negative response.
robustly studied in three randomized, placebo-controlled trials,
demonstrating safety and efficacy for both cancer and noncancer
pain (32–34). In a randomized controlled trial (RCT) in cancer pain,
which compared efficacy and side effects for IT delivery of opioids
vs. systemic delivery via conservative management, IT delivery of
opioids was superior (11).
At the present time, there is an ongoing investigative effort,
with FDA oversight, by Mallinckrodt Pharmaceuticals (St. Louis,
MO) to move IT hydromorphone from a compounded to a
branded, formally manufactured, FDA-approved product. This
endeavor involves two clinical trials. The first trial is a controlled,
two-arm, parallel group, randomized withdrawal study followed
by an openlabel single-arm safety study of hydromorphone. Fur-
ther, critical evaluations regarding sustainability and cost effec-
tiveness have been performed, with focus on sustainability and
safety.
Consensus Point 4. The use of evidence ranking is a critical piece
of the formatting of the 2016 PACC. This is the first time this impor-
tant point has been included in the PACC methods.
Consensus Point 5. In areas where evidence is strong, peer-
reviewed references are noted for the PACC recommendation.
When evidence is weak or lacking, consensus opinion is used to
make recommendations.
PAIN CARE ALGORITHM
Careful consideration of patient selection is foundational for suc-
cessful, sustainable patient care. The fact that no recommendations
were made regarding patient survival and IT therapy, or anatomic
region of pain, were deficiencies of the previous 2012 PACC. There-
fore, the PACC of 2016 is presenting evidence and consensus-based
recommendations regarding patient survival, disease process, and
medication usage for IT therapy. Attention was directed to the age
of the patient, although the contributions of age are reflected in
dosage sustainability, which is addressed elsewhere in the
recommendations.
www.neuromodulationjournal.com C 2017 International Neuromodulation Society
V Neuromodulation 2017; 20: 96–132
INTRATHECAL THERAPY BEST PRACTICES AND GUIDELINES
Algorithmic Treatment of Pain
The landscape of pain medicine and IT therapy has evolved (35).
Neuromodulation literature supports that it is more cost effective,
more efficacious, and no longer appropriate to position IT therapy
as a salvage therapy (36). The choice of neuromodulation therapy
rests on many factors, including the regional location of the pain,
type of pain, life expectancy, and malignancy, to name a few. Gener-
ally, and most significantly, IT therapy is not salvage therapy after
failure of systemic high-dose opioid-based medicines. IT drug deliv-
ery devices (IDDs) should be suggested for refractory pain, as recent-
ly defined (37):
Pain is defined as refractory, regardless of etiology, when 1)
multiple evidence-based biomedical therapies used in a clini-
cally appropriate and acceptable fashion have failed to reach
treatment goals that may include adequate pain reduction
and/or improvement in daily functioning or have resulted in
intolerable adverse effects, and when 2) psychiatric disorders
and psychosocial factors that could influence pain outcomes
have been assessed and appropriately optimized.
A proposed algorithm has been described (27,38). A modified ver-
sion of the algorithm is suggested here (Fig. 2).
Important decision-tree aspects of this algorithm deserve men-
tion. IT opioid infusion within the algorithm deserves special consid-
eration. Compared to systemic opioid delivery, IT opioids appear to
be safer based on mortality and morbidity data (39) and have fewer
diversion risks. From 2003 to 2013, opioid use disorder increased in
the United States, significantly increasing mortality, with more than
16,000 deaths reported annually (40). This has led the Centers for
Disease Control (CDC) to draft new guidelines for chronic opioid
therapy for noncancer pain. The first recommendation of those
guidelines stresses the need for nonpharmacological and nonopioid
treatments before resorting to systemic opioids. Along with neuro-
stimulation, IT therapy should play an important role in the pain
treatment continuum. The reported challenges associated with IT
therapy (41,42) have been increasingly mitigated by vigilance and
101
mindful patient selection (43).
 DEER ET AL.

Figure 3. Pain care algorithm for cancer-related pain. DRG, dorsal root ganglion; HF10, high frequency stimulation; PNfS, peripheral nerve field stimulation; PNS,
peripheral nerve stimulation; SCS, spinal cord stimulation. Green arrows indicate affirmation or positive response; red arrows signify negative response.

As can be gleaned from Figure 2, and new to descriptions of algo-
rithmic advanced pain care, IT therapy now occupies the same line
of management as neurostimulation, with important caveats. If the
patient’s condition can be treated effectively with stimulation thera-
py or IT therapy, then neurostimulation may be considered first, sec-
ondary to safety considerations (42). However, there was significant
discussion among the authors surrounding the sustainability of both
IT and neurostimulation therapies. Traditional spinal cord stimula-
tion (SCS) is an effective therapy, as clearly demonstrated in multiple
studies (44,45). However, there is a significant patient population
(nearly 30%) that becomes tolerant to the therapeutic perceived
paresthesia (46). Furthermore, Hayek and Veizi reported an explant
rate from a cohort of university hospitals of nearly 23.9% (47). New
stimulation therapies may help with salvage (48–50), although there
is a compelling argument for a role of IT therapy, certainly after SCS
failure, but also before. A significant reason for SCS explant is loss of
therapeutic coverage or development of new areas of pain since
implant. A registry study suggested approximately 8% of patients
(13/156) developed pain outside the ability of coverage of traditional
SCS at 12 months (51). Recently, data were presented at the North
American Neuromodulation Society (NANS) meeting (Las Vegas,
2015) that suggested durability of care for >6 years with IT therapy
(52). From a historical perspective, IT coverage may be influenced by
catheter position and medication selection, providing flexibility to
the clinician and the patient (8). Coverage of a common condition,
such as failed back surgery syndrome (FBSS), could potentially allow
for comparison of the efficacy of SCS with IT therapy, but no direct,
controlled studies have been conducted in the past decade.
An algorithmic hierarchy of IT therapy for cancer pain is similar.
Prognosis, likely progression of the disease into different anatomic
regions/tumor characteristics, and periprocedural imaging findings
are useful considerations for device selection. In patients with
cancer-related pain, Smith et al. demonstrated improvement in side-
effects and pain with IT therapy compared to conservative medical
management (11). Staats, in a comparative trial for AIDS and cancer
patients, showed improvement in the cancer population with IT
ziconotide (33). Furthermore, titratability and coverage/efficacy of
mechanical, nociceptive pain is less likely with SCS therapy (53).
Although coverage can be extended with traditional SCS therapy by
reprogramming, repositioning of electrodes, or adding new electro-
des, it is not accomplished to the degree or ease of IT therapy. To
that end, device selection algorithmic care is suggested in Figure 3.
End-of-life pain device selection is typically not performed if the
patient has less than three months of expected longevity. The PACC

Table 5. Recommendations for Application of Intrathecal Therapy vs. Neurostimulation by the NACC Using USPSTF Criteria.

Statement Evidence level Recommendation grade Consensus level

Intrathecal therapy should be considered within the same line as III C Moderate
neurostimulation strategies to treat noncancer-related pain.
Intrathecal therapy should be considered after neurostimulation strategies III I Strong
to treat noncancer-related pain if the pain is isolated and unlikely to
spread.
Intrathecal therapy should be considered before neurostimulation therapy III C Strong
102

for active cancer-related pain that is mechanical and likely to spread.

www.neuromodulationjournal.com C 2017 International Neuromodulation Society

V Neuromodulation 2017; 20: 96–132

Table 6. Disease Indications for Intrathecal Drug Delivery.
 Axial neck or back pain; not a surgical candidate
䊊 Multiple compression fractures
䊊 Discogenic pain
䊊 Spinal stenosis
䊊 Diffuse multiple-level spondylosis
 Failed back surgery syndrome
 Abdominal/pelvic pain
䊊 Visceral
䊊 Somatic
 Extremity pain
䊊 Radicular pain
䊊 Joint pain
 Complex regional pain syndrome (CRPS)
 Trunk pain
䊊 Postherpetic neuralgia
䊊 Post-thoracotomy syndromes
 Cancer pain, direct invasion and chemotherapy related
 Analgesic efficacy with systemic opioid delivery complicated by
intolerable side effects
recommends that in the future a more objective measure, such as
the Karnofsky Palliative Performance Scale (54) or Eastern Coopera-
tive Oncology Group Performance Status (55) scores be used. The
medication regimen (or the development of combination therapy
strategies) may be accelerated in end-of-life pain coverage. This will
be discussed with more granularity in the medication-selection
section.
Table 5 presents PACC recommendations for application of IT vs.
SCS therapy.
DISEASE-SPECIFIC INDICATIONS AND
CONSIDERATIONS
Disease-specific indications for IT therapy have been defined pre-
viously (8,27,56). A conceptual marriage of many factors contribute
to the implementation of the therapy once disease-specific indica-
tions have been fulfilled, including: survival time, opioid exposure/
sensitivity, location of pain, type of pain, medication physiochemical
properties, catheter location, pump infusion strategy, and psycho-
logical features and social support of the patient. Simply stated, IT
therapy is indicated by the Food and Drug Administration (FDA) for
moderate to severe trunk and limb pain, and intractable pain, where
more conservative therapies have failed (57,58). This includes a varie-
ty of disorders, including those highlighted in Table 6. There is
renewed interest to cover focal extremity pain with IT therapy, but
although anecdotal reports exist, literature support is lacking.
As outlined in previous versions of the PACC and other consensus
papers, it is imperative to have a clear diagnosis, an appropriate
Table 7. Cancer Patient Classifications.
Category 1 Category 2 Category 3
Patient with imminent Patient whose disease Patient with cancer in
death or life is stable or slowed, partial remission or
expectancy relatively with high likelihood cured, with residual
short, with palliation of recurrence or chronic pain
primary objective progression
www.neuromodulationjournal.com C 2017 International Neuromodulation Society
V Neuromodulation 2017; 20: 96–132
INTRATHECAL THERAPY BEST PRACTICES AND GUIDELINES
physical examination and a complete psychosocial evaluation
(which may be optional for cancer pain) before undertaking a trial or
implant. The PACC would point out that, in the face of psychological
distress from end-of-life, not having psychological help may amplify
the pain experience and compound suffering.
The cancer pain population deserves special mention, as imple-
mentation of IT therapy and medication selection, along with the
sustainability of a regimen, are largely dependent on the stage of
the disease and life expectancy (Table 7). Deer et al. have classified
cancer patients based on categories of their disease (56). The com-
plex interplay between disease indications and patient selection will
be discussed in the following section.
PATIENT-SELECTION CONSIDERATIONS
Updating the PACC documents of 2012 and mindful of patient
selection, managing patient comorbidities has been a concern since
the IDD mortality data were reported in 2009 (41). Numerous reports
have suggested best practice and careful consideration for the com-
plex interplay between disease, patient characteristics, and drugs
chosen for IT delivery. As outlined previously, consideration for all
the variables and vigilance is required, and we will address them
individually (8,9,26,27,56,60–63). Some influences of determining IT
therapy are universal, spanning all patients considered, while others
require risk stratification and more granularity (Fig. 4).
Procedural and Surgical Comorbid Disease Management
Careful consideration for comorbidities that impact wound heal-
ing (64) and the implementation of the therapy is crucial. Consider-
ations also include ability to undergo the procedure, including
anesthesia and assessment of bleeding risk. The 2014 NACC also
reported on mitigation of surgical site infection and bleeding, with
an update in review at the time of this writing (65). The PACC recom-
mendations for avoiding surgical site infections appear in Table
8 and those for anticoagulation management appear in Table 9.
Implementation of IT infusion procedurally is just one aspect of
the therapy. Decision-making regarding medication, site of delivery,
and infusion strategy must also be discussed. Cardiopulmonary
assessment is vital when determining the medication employed and
the required vigilance surrounding the implant procedure (10,27).
One of the important components of IT therapy is the fear of respira-
tory depression (27). It is important to understand the patient’s base-
line cardiopulmonary status, along with baseline medications that
may play a role in sedation and influence the CO2 response curve by
shifting it down and to the right (80,81). When considering the rela-
tive risks of oral opioids with more than 16,000 deaths per year (39),
the risks of IT agents appear to be considerably less, but no compar-
ative risk studies have been performed. With that said, it should be
emphasized that most IT infusion patients are already opiate toler-
ant, therefore, the risk of respiratory depression in such opiate-
tolerant patients is extremely low.
Disease comorbidities that may impact the influence of opioid-
based medications on respiratory drive include central or obstructive
sleep apnea, advanced age, pulmonary disease (obstructive or
restrictive), and cardiac disease (ischemic, congestive, or myopathy
related). Obstructive sleep apnea has been observed in nearly 18
million people in the United States, with physiologic responses of
hypercapnia and pulmonary hypertension. Numerous studies sug-
gest opioids negatively impact respiratory drive and may lead to
increased apnea duration (82–85). Age plays a role in the sensitivity
103
to the depressive effects of opioids (85). The American Society of
 DEER ET AL.
Figure 4. Algorithm of patient-selection characteristics. Green arrows indicate affirmation or positive response, red arrows signify negative response. Dashed
arrows signify similarity amongst groups. Cancer pain is tiered to three categories (see Table 7).
Anesthesiologists developed consensus guidelines to mitigate risk
of opioid-related respiratory depression in the acute pain popula-
tion, and although the guidelines do not translate to the chronic-
pain, opiate-tolerant patient per se, the theme is vigilance (84). Risk
stratification of patients when employing IT therapy is imperative to
provide safe pharmacologic care.
Given our current knowledge, understanding the role that the
concomitant use of central nervous system (CNS)-active medications
has on IT therapy is crucial. Careful appreciation for special popula-
tions of patients is also needed, including those with kidney or liver
disorders, along with the geriatric population, as clearance and elim-
ination (pharmacokinetics) of these CNS-active medications should
be considered. It is beyond the scope of this text to describe the
pharmacokinetics of systemic medications, but it is important to
note that potential interaction exists between IT-delivered medica-
tions and CNS-active medications. Any CNS depressant can augment
opioid-induced respiratory depression (Table 10). Creatinine kinase
(CK) levels need to be checked at baseline and even intermittently
during therapy with initiation of ziconotide, as clinical trials showed
that levels can rise to two or three times the upper limit of normal.
Causation and clinical significance of this rise is unclear, but good
clinical judgment is important and symptoms need to be carefully
evaluated if symptomatic CK elevation exists, with reduction or ces-
sation of therapy to be considered. Registry data recently suggested
improved tolerability of ziconotide when it was not infused with sys-
temic medications having a similar mechanism of action and func-
tion (86). This knowledge allows for risk stratification of dosing and
vigilance surrounding IT implementation. Paradoxically, nonopioid-
na€ıve patients may have an increased margin of safety when IT ther-
apy is initiated, although the relationship is nonlinear and overdose
remains a possibility (27).
Previous Systemic Opioid Use
Discussion of patient selection requires considering all of the
aforementioned factors of concurrent medications, disease states,
prior treatments, and device implementation. Notwithstanding, the
concept of bulk flow within the IT space is limited (8,20,90–92), with
potentially very little spread from the distal end of the catheter with
104
the CSF. It has long been appreciated that the supratentorial effects
www.neuromodulationjournal.com C 2017 International Neuromodulation Society
V Neuromodulation 2017; 20: 96–132
reported surrounding delayed respiratory depression with acute IT
opiate delivery are different from those with chronic infusion. Chron-
ic infusion is associated with slow, low volume, and low kinetic ener-
gy delivery wherein local rostrocaudal IT distribution is limited.
Redistribution of CNS-active, IT-delivered drugs out of the IT space is
likely the culprit for respiratory effects (80,81). This concept is corrob-
orated by various recent works (8,27,55).
This new understanding has been validated in animal models,
with evidence suggesting the same mechanism in humans (93). Sim-
ilarly, previous exposure is relevant. More patients are taking system-
ic opioid medications than ever before in the United States, with
increased mortality. Diversion and misuse contribute to the associat-
ed morbidity and mortality (40,94–96). The California State Medical
Board has created guidelines for systemic opioids to be less than 80
morphine equivalents (MEs), while others advocate for less than 100
or 120 MEs (95,97). The CDC recently recommended chronic mainte-
nance doses of systemic opioids of 50 ME or less, with extra vigi-
lance and risk assessment at >90 ME. The history of systemic
exposure is important when considering IT therapy, not just for cal-
culating the total amount of systemic opioids, but also for consider-
ing the impact of IT opioid dose escalation, as failure of systemic
opioids where side effects are not an excluding factor suggests fail-
ure of opioids intrathecally. Preclinical work has shown cross-
tolerance between IT and systemic morphine (96).
In one IT opioid study of patients with chronic pain, during the
washout period before randomization to another IT drug, the major-
ity of patients were able to wean IT opioid medications but with
average systemic supplementation of near 300 MEs per day (32).
The patients had visual analogue scale of pain intensity (VASPI)
scores of >80 mm, suggesting failing IT therapy. If patients are
maintained on large amounts of systemic opioids before starting IT
therapy, the likelihood of monotherapy failure seems to be higher
(12,14,15,38,99,100). This concept supports the approach of weaning
or reducing systemic opioids prior to initiating IT therapy.
This potency shift is not predictable based on previous models for
systemic to IT conversion (27), and there are reports of tolerance
reversal in as little as a week off IT agents (42). Veizi et al. reported a
reduction of dose escalation of nearly three times when employing
a local anesthetic (99). Many authors recommend weaning strategies
 INTRATHECAL THERAPY BEST PRACTICES AND GUIDELINES

Table 8. Recommendations for the Avoidance of Surgical Site Infections as Recommended by the PACC.

Recommended practice Origin of recommended References and

practice§ comment
Preoperative practices

Utilization of preoperative antibiotics for trials
Utilization of preoperative weight-based antibiotic dosing for
trials
Utilization of preoperative antibiotics for implants
Utilization of preoperative weight-based antibiotic dosing for
implants
Utilization of preoperative nasal culture and application of
mupirocin to prevent Staphylococcus aureus surgical site
infections
Appropriate preoperative timing (within 1 hour prior to
surgical incision excluding vancomycin) of prophylactic
antimicrobial administration for implants
Hair removal (when required) with electric clippers
immediately before the surgical procedure
CDC IA and NICE Bowater et al. (66) demonstrated that antibiotic
prophylaxis is effective for reducing the risk of wound
infection for all types of surgery.
CDC IA and NICE Weight-based dosing of antibiotics is required to achieve
therapeutically effective drug concentrations (67,68).
CDC IA and NICE Bowater et al. (66) demonstrated that antibiotic
prophylaxis is effective for reducing the risk of wound
infection for all types of surgery.
CDC IA Weight-based dosing of antibiotics is required to achieve
therapeutically effective drug concentrations (66,68).
CDC 1A Universal decolonization seems associated with a low risk
of mupirocin resistance in Staphylococcus aureus (69).
Mupirocin and antiseptic body wash may reduce
superficial but not deep surgical site infections (70).
Mupirocin decolonization is effective in a high-risk popula-
tion (71), however, compliance is low and resistance
occurs (72,73). Nasal povidone-iodine may be considered
as an alternative (74).
CDC IA, NICE, SCIP
CDC IA and NICE

Intraoperative practices
Utilization of double gloving CDC II and NICE Tanner and Parkinson (75) concluded that the addition of
a second pair of surgical gloves reduces perforations to
the innermost gloves. Although there is insufficient
evidence that double gloving reduces the risk of SSIs,
NICE recommends wearing two pairs of sterile gloves
when there is a high risk of glove perforation and the
consequences of contamination may be serious (76).
Utilization of chlorhexidine gluconate for preoperative skin CDC IB and NICE CDC and NICE recommend the use of an appropriate
antiseptic agent antiseptic agent (povidone–iodine or chlorhexidine-
containing products). Darouiche et al. (77) demonstrat-
ed that preoperative skin preparation with
chlorhexidine-alcohol is superior to povidone–iodine for
preventing SSI.

Postoperative practices
Application of an occlusive dressing following a trial CDC IB and NICE CDC recommends applying a sterile dressing for 24–48
hours postoperatively (category IB). NICE recommends
interactive dressings. Hutchison and McGuckin
demonstrated lower rates of infection with occlusive
dressings (78).
Understanding maximum time criterion for defining a deep CDC Infection occurs within one year if an implant is in place
surgical site infection of an implantable device (one year) and the infection appears related to the operation.
No continuation of antibiotics into the postoperative period SCIP SCIP recommends the discontinuation of antibiotics
for trials beyond 24 hours* within 24 hours after surgery.
No continuation of antibiotics into the postoperative period SCIP SCIP recommends the discontinuation of antibiotics
for implants beyond 24 hours† within 24 hours after surgery.
CDC, centers for disease control; NICE, National Institute for Health and Care Excellence; SCIP, Surgical Care Improvement Project of the Joint
Commission; SSIs, surgical site infections.
*Examination of survey questions 22 and 23.
†
Examination of survey questions 24 and 25.
§
CDC recommendations. IA: Strongly recommended for implementation and supported by well-designed experimental, clinical, or epidemiological stud-
ies. IB: Strongly recommended for implementation and supported by some experimental, clinical, or epidemiological studies and strong theoretical ratio-
nale. II: Suggested for implementation and supported by suggestive clinical or epidemiological studies or theoretical rationale.
105

www.neuromodulationjournal.com C 2017 International Neuromodulation Society

V Neuromodulation 2017; 20: 96–132
 DEER ET AL.

Table 9. Anticoagulation Management Practices for Intrathecal Therapy as Recommended by the PACC.

Anticoagulant Recommendation for trial Recommendation for permanent implant

Warfarin Discontinue five to seven days before, INR <1.5; if Discontinue five to seven days before, INR
bridging required, refer to bridging medication; <1.5; if bridging required, refer to bridging
continue cessation during duration of trial, medication; resume 24 hours postoperatively.
resume 24 hours following trial catheter removal.
Enoxaparin (LMWH) Hold therapeutic dose of LMWH 24 hours before Hold therapeutic dose of LMWH 24 hours
procedure; hold for duration of trial; resume 24 before procedure; resume 24 hours
hours following catheter removal. following surgery
Clopidogrel (ADP receptor antagonists) High-risk patients for cardiac events—discontinue High-risk patients for cardiac events-discontinue
at least five days before; low risk seven to ten at least five days before; low risk seven to
days before; hold for duration of trial; resume ten days before; resume 24 hours following
24 hours following catheter removal. surgery.
Prasugrel (ADP receptor antagonist) Discontinue seven to ten days prior to procedure, Discontinue seven to ten days prior to
hold for duration of trial, resume 24 hours procedure, hold for duration of trial,
following catheter removal. resume 24 hours following lead removal.
Ticlopidine (ADP receptor antagonists) Discontinue 14 days prior to procedure, hold for Discontinue 14 days prior to procedure;
duration of trial, resume 24 hours following resume 24 hours following surgery.
catheter removal.
Abciximab, eptifibatide, tirofiban Discontinue for three days prior to procedure, Discontinue for three days prior to procedure,
(platelet GPIIb/IIIa receptors) hold for duration of trial, restart 24 hours hold for duration of trial, restart 24 hours
following catheter removal.‡ following the surgery.‡
Dipyridamole, aggrenox (aspirin/ Discontinue seven days prior to procedure, Discontinue for seven days prior to procedure,
dipyridamole) (phosphodiesterase hold for duration of trial, restart 24 hours hold for duration of trial, restart 24 hours
inhibitors) following catheter removal.§ following the surgery.§
Naproxen, ketorolac, ibuprofen, Discontinue seven days prior to procedure, Discontinue seven days prior to procedure,
etodolac, etc. (nonsteroidal hold for duration of trial, reinitiate 24 hours hold for duration of trial, reinitiate 24 hours
anti-inflammatory drugs)§ following catheter removal. following the surgery
Aspirin§ Discontinue seven days prior to procedure, Discontinue seven days prior to procedure,
hold for duration of trial, reinitiate 24 hours hold for duration of trial, reinitiate 24 hours
following catheter removal. following surgery.
Herbals (ginseng, ginko, garlic) Discontinue seven days prior to the procedure, Discontinue seven days prior to the procedure;
hold for duration of trial, reinitiate 24 hours reinitiate 24 hours following surgery.
following catheter removal.
Dabigatran etexilate mesylate, Discontinue five days prior to procedure, hold Discontinue five days prior to procedure,
rivaroxaban (direct thrombin for duration of trial, reinitiate 24 hours hold for duration of trial, reinitiate 24 hours
inhibitors) following catheter removal. following surgery.
Heparin IV* On a case-by-case basis On a case-by-case basis
Heparin SQ† On a case-by-case basis On a case-by-case basis
INR, international normalized ratio; ADP, adenosine diphosphate; LMWH, low molecular weight heparin; NA, not available; IV, intravenous; SQ,
subcutaneous.
*Requires inpatient hospitalization and monitoring, suggesting a special need or indication for neurostimulation and should be assessed on case-by-case
basis.
†
Peaks at 2–4 hours after administration, typically thrombotic prophylaxis as inpatient and may require platelet assessment if more than four-day dosing.
Please refer to American Society for Regional Anesthesia guidelines and determine on a case- by-case basis.
‡
Typically contraindicated four weeks following surgery. If reinitiated, careful follow-up and vigilance is suggested (79).
§
Current recommendations (79) suggest variable stoppage is necessary based on clinical context, and on the specific half-life of the nonsteroidal anti-
inflammatory in question. The half-life determines the time required for discontinuation in order to limit the drug’s effect on platelet function.

(101), with the core concept of improving patient selection by mov- requirement is somewhat obscured with the recently described
ing away from salvaging high-dose systemic opioid failure. The pre- weaning strategies, but the theme is suggested (13).
dictive value of systemic dose requirement to IT opioid dose The concept of limiting systemic opioid exposure to less than 100
MEs has already been proposed (1,38), with the CDC recommending
less than 50 ME for noncancer related chronic pain and increased
Table 10. Central Nervous System-Active Medication Classes That May vigilance above 90 MEs (102). Although no specific data comparing
Interact With Intrathecal Opioids (86–88). prior systemic dose to IT dose exist, enough prospective and retro-
spective data suggest it may be beneficial as a tool for IT dose esca-
Benzodiazepines
lation prediction (13,35). Furthermore, there does not seem to be a
Antidepressants
gender difference with opioid dose, although in one study women
Anticonvulsants
Muscle relaxants were more likely to continue to receive oral opioid adjuvants (97).
Hatheway et al. described the use of patient-controlled boluses to
106

EtOH consumption
minimize systemic opioid use with IT therapy (103).

www.neuromodulationjournal.com C 2017 International Neuromodulation Society

V Neuromodulation 2017; 20: 96–132

Table 11. Recommendations for Patient Selection Criteria for Intrathecal Therapy by the PACC Using USPSTF Criteria.
Statement
Patients with comorbidities that negatively affect cardiopulmonary function
need increased vigilance when instituting intrathecal opioid therapy.
Localized pain can be adequately covered with intrathecal therapy.
Diffuse pain can be adequately treated with intrathecal therapy.
Global pain can be adequately treated with intrathecal therapy.
Intrathecal therapy should not be used as salvage therapy for failing
systemic opioids.*
*Patients who are weaned prior to the trial have a higher likelihood of sustained success (15,101). Different titration schedules have been recommended.
Sustainability of Intrathecal Opioid Therapy
Dose escalation with chronic IT opiates has been a cause for con-
cern. An early retrospective study noted that in cancer patients
receiving IT infusion of morphine, for periods in excess of three
months after initial stabilization (4.8 mg/d), 48% showed less than a
twofold increase in dose by three months (104). In 52 cancer
patients, initial dose (3.8 mg/d) rose by a factor of 2.5 at four months
(105). In 50 cancer patients, receiving infusion for a mean of 142
days (7–584 days), the mean IT morphine starting dose, 2.5 mg/day,
increased to a mean final dose of 9 mg/day (106). In a series of
chronic noncancer pain patients (n 5 88), mean IT morphine dose
increased from 10 mg/day at 6 months to 15 mg/day at 36 months
after initiation of therapy (107). In another study, presence of neuro-
pathic pain seemed to be strongly predictive of IT dose escalation
(108). Seemingly, patients less than 50 years old have a greater
chance of opioid dose escalation intrathecally compared to patients
older than age 50 (12). Mitigating strategies include proper patient
selection, minimizing pre-IT therapy opioid dose, proper localization
of pain, and congruent catheter placement. Local anesthetics, such
as bupivacaine have been shown to have synergistic interaction
with opiates (109,110). Veizi et al. reported dose escalation mitiga-
tion with concurrent use of bupivacaine, and although this is off-
label, it appears logical and should be considered (99).
Psychological Assessment and Social Support Evaluation
Previous consensus guidelines, best practices, and previous PACC
iterations have all commented on psychological assessment of
patients for IT therapy (8,21,22,27). There has been little update on
patient selection surrounding this topic. Category 1 cancer patients
do not need a robust psychological screen, as palliation is the goal
of IT therapy, however, they may benefit from counseling regarding
death, dying and chronic illness. For all others, clear descriptions of
developing a partnership and setting expectations between patient
and treatment team is suggested in the literature. Importantly, zico-
notide is contraindicated in patients with a history of psychosis, and
alternative IT medications should be used (111).
Pain Characteristics: Regional vs. Diffuse vs. Global
There is little doubt that CSF flow dynamics and the pharmacoki-
netic profiles of IT agents are better understood now than ever
before. Little bulk flow exists (90–92). Rostral spread within the IT
space is limited from the catheter tip. Although there is limited evi-
dence to support catheter placement congruent with the dermato-
mal area of pain, consensus remains high that this is a crucial
component to sustained IT care. Similarly, placement of the catheter
dorsal to the spinal cord delivers medication closer to the dorsal ele-
ments of the spinal cord. Localized pain examples include the band-
like dermatomal challenges of postherpetic neuralgia or unilateral
www.neuromodulationjournal.com C 2017 International Neuromodulation Society
V Neuromodulation 2017; 20: 96–132
INTRATHECAL THERAPY BEST PRACTICES AND GUIDELINES
Evidence level Recommendation grade Consensus level
III C High
II B Strong
III C Moderate
III D Moderate
II B Moderate
precise abdominal pain or axial back pain from vertebral compres-
sion fracture. Diffuse pain refers to a whole extremity involvement,
back or leg discomfort, abdominal pain encompassing more than
one quadrant, and so on. Global pain is total body pain.
The PACC recommendations for patient selection appear in
Table 11.
Consensus Point 6. Patients with comorbidities that negatively
affect cardiopulmonary function need increased vigilance when
instituting intrathecal opioid therapy.
Consensus Point 7. Localized, diffuse and global pain can be ade-
quately treated with intrathecal therapy. The evidence for global pain
treatment is less well defined and should be approached cautiously.
Consensus Point 8. Intrathecal therapies should be used at an
appropriate time in the algorithm and not as a salvage treatment.
MEDICATION-SELECTION RECOMMENDATIONS
AND CONSIDERATIONS
Since the publication of the PACC reports of 2012, more energy by
these authors has focused on patient selection, procedure standardi-
zation, and infusion therapies compared to new medications. The
FDA has approved ziconotide and morphine for IT infusion for the
treatment of pain. Hydromorphone from Mallinckrodt plc. is under-
going clinical trial for potential IT labeling. Notwithstanding, the
PACC of 2012 provided a framework to determine which IT medica-
tions to use when differentiating the patient’s pain as either nocicep-
tive or neuropathic or mixed pain. This framework of understanding
has been expanded in 2016. Off-label monotherapy or combination
therapy should be considered after failure of FDA-approved medica-
tions or when these medications are contraindicated. The question
of which medication or combination of medications to use is impor-
tant and is based on safety. The current PACC algorithms were creat-
ed to help guide clinicians in the safe and effective use of IT therapy;
however, physicians should use their own best clinical judgment in
making treatment decisions for their patients.
Consensus Point 9. Off-label drug monotherapy or combination
therapy is not recommended until FDA-approved drugs are tried
and failed or are contraindicated. In cancer pain, the on-label drugs
can be used during the trial phase. If the results are not acceptable
due to lack of efficacy or side effects, an admixture with bupivacaine
or the primary use of fentanyl is supported by our consensus.
Consensus Point 10. The PACC algorithms are based on improv-
ing safety and efficacy in clinical practice, which includes the use of
107
off-label drugs.
 DEER ET AL.

Table 12. Cancer or Other Terminal Condition-Related Pain With Localized Nociceptive or Neuropathic Pain.

Line 1A Ziconotide Morphine

Line 1B Fentanyl Morphine or fentanyl 1
bupivacaine
Line 2 Hydromorphone Hydromorphone 1 Hydromorphone or Morphine or
bupivacaine fentanyl or morphine 1 hydromorphone or
clonidine fentanyl 1 ziconotide
Line 3 Hydromorphone or Ziconotide 1 Ziconotide 1 clonidine Hydromorphone or Sufentanil
morphine or fentanyl 1 bupivacaine morphine or fentanyl 1
bupivacaine 1 clonidine bupivacaine 1 ziconotide
Line 4 Sufentanil 1 ziconotide Sufentanil 1 Baclofen Sufentanil 1 clonidine Bupivacaine 1 Bupivacaine 1
bupivacaine clonidine 1 clonidine
ziconotide
Line 5 Sufentanil 1 bupivacaine 1 clonidine
Line 6 Opioid* 1 bupivacaine 1 clonidine 1 adjuvants†
*Opioid (all known intrathecal opioids).
†
Adjuvants include midazolam, ketamine, octreotide.

Consensus Point 11. The algorithms are based on evidence and medications and is greatly dependent on catheter location. Neuro-
consensus on safety. The patient’s physician and good clinical judg- pathic pain generally responds to ziconotide, opioid plus local anes-
ment should guide individual patient care. thetic, local anesthetic alone, clonidine plus opioid, and clonidine
Considerations of the highlighted variables include patient diag- alone. Nociceptive pain generally responds to opioid, ziconotide,
noses and expected patient survival time (11), sustainability of the IT opioid plus local anesthetic, and local anesthetic alone. Manipulative
regimen (12,13), previous exposure to opioids (14,15), location of variables that can increase drug spread include the physiochemical
pain (diffuse vs. localized vs. global), type of pain (nociceptive, neu- properties of the drug, kinetics of the injectate, and the volume
ropathic, or mixed), the physiochemical properties of lipid solubility delivered (16,90). Increased concentration and daily doses of opioids
of the IT drugs (16,17), CSF dynamics and pharmacokinetics (18–20), (except fentanyl) are associated with granuloma formation, and care
catheter location (20), pump and catheter characteristics, kinetics of should be taken when dose escalation is rapid or doses and concen-
the IT infusate (20), and psychological status (21–23) of the patient trations are known to be reaching levels associated with granuloma
with chronic pain. formation (8,10). The possibility of granuloma formation should be
considered when employing opioid-based medications.
The PACC 2016 recommendations allow integration of the
INTRATHECAL THERAPY IN NEUROPATHIC applied 2012 PACC nociceptive and neuropathic localized and dif-
fuse pain recommendations for cancer or terminal illness
AND NOCICEPTIVE PAIN STATES (Tables 12–15), and for noncancer pain (Tables 16–19). For medica-
tion selection within the tiered recommendations, it is important to
Employing the familiar nociceptive and neuropathic pain classifi-
consider age, type of pain, and anticipated duration of therapy. In
cation for medication selection as a framework, the reorganization
cancer pain, the evidence would suggest combination therapy
of medication selection is based on many factors, including survival
might be warranted as a first- line strategy, which is a different rec-
expectation, age, previous exposure to opioids, location of pain,
ommendation from that for treatment of noncancer pain. If the
type of pain, and catheter location. The pharmacokinetics of the IT
patient responds to morphine or ziconotide as single medications
medications employed (112) point toward a potentially greater
during a trial, we would still recommend that on-label drugs be
spread with multiple bolus delivery compared to continuous
used initially. This distinction is highlighted by the evidence-
infusion. weighted tier system designation of line 1A and 1B. Line 1A repre-
Nociceptive and neuropathic pain, regardless of age, expected sents medication with level I evidence. Furthermore, these medica-
survival, or pain location, may respond to a combination of tions are FDA approved, with the intent of honoring the evidence
available. However, if the patient does not respond to on-label
Table 13. Cancer or Other Terminal Condition-Related Pain With
monotherapy during the trial phase, then fentanyl and combination
Localized Pain: Evidence Level, Recommendation Strength, and Consen- therapies, including admixtures with bupivacaine, are supported by
sus Level. the consensus. Baclofen is FDA approved for spasticity and is some-
times helpful in managing pain associated with spasticity.
Tier Evidence level Recommendation grade Consensus level Titratability for patients with cancer pain (Category 1 and 2, see
Table 7) is extremely important. There is significant evidence that
Line 1A I A Strong suggests opioid 6 bupivacaine is helpful in this population
Line 1B II-1 B Strong
(113,114). Careful attention should be made by caregivers to escala-
Line 2 II-3 B Strong
tion of dose or concentration above certain recommended levels (8).
Line 3 III C Moderate
Line 4 III I Weak
The tiered selection of the recommended medications was based
Line 5 III I Weak on levels of evidence surrounding the safety and efficacy of the med-
ication provided. Ziconotide monotherapy and opioid 1 bupivacaine
108

Line 6 III I Weak

share level I evidence for their use in the cancer population

www.neuromodulationjournal.com C 2017 International Neuromodulation Society

V Neuromodulation 2017; 20: 96–132
 INTRATHECAL THERAPY BEST PRACTICES AND GUIDELINES

Table 14. Cancer or Other Terminal Condition-Related Pain With Diffuse Nociceptive or Neuropathic Pain.

Line 1A Ziconotide Morphine

Line 1B Hydromorphone Morphine or hydromorphone 1 bupivacaine
Line 2 Hydromorphone or Morphine or hydromorphone 1 ziconotide
morphine 1 clonidine
Line 3 Hydromorphone or Ziconotide 1 bupivacaine Ziconotide 1 Hydromorphone or Sufentanil
morphine or fentanyl 1 clonidine morphine or fentanyl 1
bupivacaine 1 clonidine bupivacaine 1 ziconotide
Line 4 Sufentanil 1 ziconotide Baclofen Sufentanil 1 Sufentanil 1 clonidine Bupivacaine 1 clonidine 1 Bupivacaine 1 clonidine
bupivacaine ziconotide
Line 5 Sufentanil 1 bupivacaine 1 Sufentanil 1 Sufentanil 1 clonidine 1
clonidine bupivacaine 1 ziconotide
ziconotide
Line 6 Opioid* 1 bupivacaine 1 clonidine 1 adjuvants†
*Opioid (all known intrathecal opioids).
†
Adjuvants include midazolam, ketamine, octreotide.

(11,33). Little evidence suggests that ziconotide is helpful in com-
bination with opioids to manage cancer pain. No prospective, con-
trolled studies of ziconotide plus opioids have been performed.
Stability of admixtures becomes a concern when multiple medica-
tions are used for combination therapy (115). There is also concern
that drug mixing may increase permeability of medications into
the pump rotor, resulting in corrosion and pump failure (58).
Reports of ziconotide plus bupivacaine have little supportive evi-
dence. For palliative reasons in the cancer population more com-
binations of medicines are trialed, and Tier 6 was added to the
PACC 2012 algorithm. When contemplating higher-tiered recom-
mendations, it is crucial to consider the category of cancer pain
that the patient has (see Table 7).
Consensus Point 12. The disease process should be considered
when making decisions on algorithms for patient care.
Consensus Point 13. The stage of cancer and survival time
should be considered when considering drug titration.
Recommendations for medication selection for noncancer local-
ized pain need to be approached mindfully, and age and pain type
should be carefully considered. As can be seen in the noncancer-
pain tiered algorithm, and assuming that the catheter location is
congruent to the painful area, medication recommendations are
based on the physiochemical properties of the drug. Built into the
algorithms are pathways for diffuse or local, nociceptive, and neuro-
pathic pain types. The evidence behind Tiers 2–5 (or 6 for cancer or
end-of-life pain) is largely dependent on experience/consensus from
the PACC members, with graded strength of consensus.
Table 15. Cancer or Other Terminal Condition-Related Pain With
Diffuse Pain: Evidence Level, Recommendation Strength, and Consensus
Level.
Tier Evidence level Recommendation grade Consensus level
Line 1A I B Strong
Line 1B II B Moderate
Line 2 II-3 B Strong
Line 3 III C Moderate
Line 4 III I Weak
Line 5 III I Weak
Consensus Point 14. Ziconotide has strong clinical evidence for
efficacy.
Consensus Point 15. There are no cases of death from ziconotide
overdose and no granuloma formation has been reported.
Consensus Point 16. Unless contraindicated, ziconotide should
be the first drug selected in the population of noncancer patients
discussed in this consensus.
RECOMMENDED STARTING DOSAGES
Starting dosage ranges of IT medications recommended by the
PACC panel have not changed since the PACC of 2012 (Tables 20
and 21). These doses assume chronic continuous infusion. Bolus
strategies have been reported (1,116), but there are limited data to
support widespread adoption. IT dosing studies with bolus-only or
bolus-weighted infusion strategies are presently ongoing (35).
Appropriate starting opioid dosages may vary according to the
patient’s baseline oral intake at the time IT therapy is initiated, and it
is suggested that patients be stratified by risk regarding cardiopul-
monary depression and site of service initiation. Conservative initia-
tion dosing strategies are recommended.
It is important to consider morbidity and mortality data of IT deliv-
ery when gauging appropriate starting doses. Coffey et al. reported,
from device registration and Social Security death master file analy-
ses, an IT opioid therapy mortality rate of 0.088% at three days after
implantation, 0.39% at one month, and 3.89% at one year (41). All
patients were initiated on an opioid dose of >0.5 mg/day. It is also
important to contrast IT drug delivery mortality with that for pre-
scription drug overdose death rates, which quadrupled between
2000 and 2014, from 1.5 to 5.9 deaths per 100,000 people (117).
Site of service for trialing and dosing of IT therapy is an important
issue when considering morbidity and mortality. This article is
accompanied by a PACC recommendation article that pays exclusive
attention to trialing (118), where similar clarity is required. Conserva-
tive dosing, regardless of patient risk assessment, is highly recom-
mended in both articles. Risk assessment not only includes the
biologic disease indications and patient-selection criteria aforemen-
tioned, but also the nonbiologic site of service where trialing and
dosing changes occur. It was recommended in the previous PACC to
perform opioid trials and initiate monotherapy or combination opi-
109

Line 6 III I Weak

oid therapy followed by a 23-hour observation period, which

www.neuromodulationjournal.com C 2017 International Neuromodulation Society

V Neuromodulation 2017; 20: 96–132
 DEER ET AL.

Table 16. Noncancer-Related Pain With Localized Nociceptive or Neuropathic Pain.

Line 1A Ziconotide Morphine

Line 1B Fentanyl Fentanyl 1 bupivacaine
Line 2 Fentanyl 1 clonidine Hydromorphone or morphine 1 Fentanyl 1 bupivacaine 1 clonidine Bupivacaine
bupivacaine
Line 3 Fentanyl 1 Morphine or hydromorphone 1 Ziconotide 1 clonidine or Bupivacaine 1
ziconotide 1 clonidine bupivacaine or both clonidine
bupivacaine
Line 4 Sufentanil 1 bupivacaine or clonidine Baclofen Bupivacaine 1 clonidine 1 ziconotide
Line 5 Sufentanil 1 bupivacaine 1 clonidine Sufentanil 1 ziconotide

typically requires hospital admission. In review of publications and
consensus opinion since that publication, we have modified our rec-
ommendation so that the site of service for IT therapy be based on
several mitigating factors.
In the perioperative setting for trials, the safe IT dose of morphine
was determined in a meta-analysis to be near 0.075 and 0.15 mg for
a single IT bolus, although the disparity of the definition of clinically
significant respiratory depression muddies the ability to determine
the exact incidence (119). In a retrospective comparison study of
parturients in labor investigating opioid-related side effects of mor-
phine or bupivacaine, the incidence of opioid-related side effects
did not statistically differ (pruritus, nausea, vomiting, respiratory
depression). None of the 114 patients in this study had respiratory
depression at doses of 0.04 mg hydromorphone or 0.1 mg morphine
when observed for no less than 24 hours (120). In an assessment of
1524 postoperative patients, one patient who received a single-
bolus IT dose had less than ten breaths per minute (121). In a pro-
spective study comparing morphine and fentanyl in the parturient
population for cesarean delivery, there were no respiratory depres-
sion events defined as less than ten breaths per minute (122). In a
comparison of spinal analgesia for transurethral surgical procedures,
in patients who received 25 mcg of fentanyl intrathecally, no patient
(n 5 20) experienced respiratory depression or complications (122).
Although not specific to the chronic pain patient undergoing IT trial
for candidacy of IT chronic delivery, evidence suggests low-dose opi-
oid trialing is safe in the outpatient setting.
Does the safety profile described during IT trialing translate to the
initiation of therapy as an outpatient? The mortality data for IT thera-
py includes data from implantation, revision, and refill of the device
(41). There are no data suggesting safety or danger of IT opioid initi-
ation in an outpatient setting. However, it is suggested by expert
consensus that the 24-hour initiating IT dose be half of the effica-
cious/successful trialed IT opioid dose. For ziconotide, the previous
recommendation of a 12-hour observation period after initiation has
Table 17. Noncancer-Related Pain With Localized Nociceptive or Neu-
ropathic Pain: Evidence Level, Recommendation Strength, and Consen-
sus Level.
Tier Evidence level Recommendation grade Consensus level
Line 1A I A Strong
Line 1B II-3 B Strong
Line 2 II-3 B Strong
Line 3 III C Moderate
Line 4 III I Weak
Line 5 III I Weak
110
been revised, by consensus, to six hours as long as there are no
signs of neurologic dysfunction prior to initiation. The risk of mor-
phine overdose applies when using higher initial drug concentra-
tions of morphine. This is important since the FDA considers drug
dilution as an off-label use of drug. In settings where the initial drug
concentrations create the need for a starting dose outside of the
PACC recommendations, an overnight admission is advised.
No evidence suggests superiority of one trialing method over
another, which includes duration of trial (15,124). However, inpatient
catheter trials offer the flexibility to trial different intrathecal medica-
tions and regimens following one dural puncture, and may be help-
ful in the complex patient.
Special comment needs to be made here regarding patient-
controlled bolus administration and IT opioid rotation or a medica-
tion switch. The 2012 PACC recommendations suggested that
patient-controlled dosages be 5–20% of the total daily dosage. In a
retrospective review from the Cleveland Clinic, up to 30% of the
24-hour dose could be administered safely during each patient-
activated dose, up to four times daily (125). This represents a signifi-
cantly larger incremental increase of the 24-hour dose, suggesting
that more prospective or retrospective data are required. In another
unpublished study from the Cleveland Clinic, patient-controlled IT
analgesia proved to be cost-effective, paying for the IT device in
eight months (126).
Consensus Point 17. The initiating dose of intrathecal opioids
and ziconotide should be as low as reasonably expected to provide
analgesia.
Consensus Point 18. The initiating dose of intrathecal opioids
and ziconotide delivered continuously should be 50% or less of the
dose used during bolus trialing.
Consensus Point 19. The PACC recommends that the primary
medication be weaned and discontinued when converting medica-
tions from one single medication to a different single medication in
the algorithm. The use of ziconotide and bupivacaine do not have
risk of withdrawal and weaning is not needed. The abrupt stopping
of an opioid is not recommended.
Consensus Point 20. The PACC recommends careful attention to
side effects when adding any adjuvant drug to a primary drug.
Consensus Point 21. Medications with significant withdrawal syn-
dromes, including clonidine and baclofen, require rescue strategies
in the event of abrupt cessation or interruption in intrathecal
delivery.
Maximal recommended daily doses based on preclinical studies,
animal toxicity studies and consensus were published in the 2012

Line 6 III I Weak

PACC. Preclinical work with fentanyl and sufentanil used single

www.neuromodulationjournal.com C 2017 International Neuromodulation Society

V Neuromodulation 2017; 20: 96–132
 INTRATHECAL THERAPY BEST PRACTICES AND GUIDELINES

Table 18. Noncancer-Related Pain With Diffuse Nociceptive or Neuropathic Pain.

Line 1A Morphine Ziconotide*

Line 1B Hydromorphone Morphine or hydromorphone 1
bupivacaine
Line 3 Hydromorphone or Fentanyl 1 bupivacaine Ziconotide 1 morphine or
morphine 1 clonidine hydromorphone
Line 4 Hydromorphone or Fentanyl 1 ziconotide Sufentanil 1 bupivacaine or clonidine Ziconotide 1 clonidine or
morphine 1 bupivacaine 1 bupivacaine or both
clonidine
Line 5 Fentanyl or sufentanil 1 bupivacaine 1 clonidine Sufentanil 1 ziconotide Baclofen
Line 6 Opioid 1 ziconotide 1 bupivacaine or clonidine
*Ziconotide should be first choice in patients with >120 morphine equivalents or fast systemic dose escalation, in the absence of history of psychosis.

bolus, multiple daily boluses, or continuous infusions across many
animal models (127–133). It appears that high concentrations of fen-
tanyl and sufentanil are tolerated, exclusive of a single injection
sheep study (133). Very limited continuous infusion studies exist.
Although some data suggest that 50 mcg/mL has limited data sup-
porting safety, the 2000 mcg/mL study also suggests that neurotox-
icity may not be a concern, with conflicting reports with fentanyl up
to concentrations of 5000 mcg/mL (127,128).
Although these animal toxicity studies do support the safety of
high concentrations of fentanyl and sufentanil, there was concern
among PACC members about the phraseology of the “no known
upper limit” recommendation. Those expressed concerns included:
1) high daily doses may lead to opioid- induced hyperalgesia; and 2)
high plasma levels of the medication from very high IT doses may
limit the advantages of targeted drug delivery to the spinal dorsal
horn. The majority of PACC members agreed that in most clinical set-
tings patients would not benefit from daily doses higher than 1000
mcg of fentanyl or 500 mcg of sufentanil. Members also agreed that
the lowest possible concentration should be used, as human data
are limited. Therefore, the consensus was that a more conservative
approach to dosing of both fentanyl and sufentanil should be taken.
The consensus panel did note the experience of several members
suggesting that escalation of IT fentanyl doses was often associated
with diminishing returns. Several have observed clinical responses
consistent with hyperalgesia and lack of clinical efficacy in dose
ranges above 1000 mcg per day of IT fentanyl. Further systemic
absorption of this highly lipophilic opioid may approach systemic lev-
els seen with transdermal systemic applications (>3 ng/dL) as higher
IT doses are utilized. While there is no conclusive data to guide the
panel, the PACC does suggest strong reevaluation and consideration
of other approaches as doses cross the 1000 mcg per day dosing
threshold. Clearly clinicians have utilized doses above the 1000 mcg
Table 19. Noncancer-Related Pain With Diffuse Nociceptive or Neuro-
pathic Pain: Evidence Level, Recommendation Strength, and Consensus
Level.
Tier Evidence level Recommendation grade Consensus level
Line 1A I A Strong
Line 1B II B Strong
Line 2 II-3 B Strong
Line 3 III C Moderate
Line 4 III I Weak
Line 5 III I Weak
per day level safely and possibly with efficacy, however, the panel
does recommend this dosage currently as a reevaluation milestone.
Similarly, evidence suggests that the much higher doses of bupi-
vacaine are well tolerated and average concentrations are reported
in many studies (99,134,135). For all of the other drugs, the recom-
mendations established in 2012 are still supported in the evidence,
and we endorse the same dosing recommendations in 2016
(Table 22).
Consensus Point 22. Before proceeding with aggressive dose
titration above 1000 mcg per day of IT fentanyl, clinicians should
closely monitor the outcome of each dose increase and, if efficacy is
not being established, consider dose reduction with consideration
of intrathecal tolerance and hyperalgesia.
VARIABLES AFFECTING CHRONIC
INTRATHECAL THERAPY
Spinal Anatomy and CSF Dynamics Relevant to IT Drug
Delivery
Meninges are morphologically and physiologically implicated in
mechanical, immunologic, trophic, metabolic and thermal protec-
tion of the brain and spinal cord. In relation to spinal drug delivery,
the spinal meninges represent the main barrier to the transfer of
drugs between the CSF and the spinal cord. Therefore, it is necessary
to know if any of the meninges cause resistance or limitation to the
free circulation of CSF, presenting a barrier or compartmental limita-
tion. The spinal dural sac contains the subarachnoid space with the
trabecular arachnoid, the pia mater, and the subpial tissue. Drugs
must cross all of these structures before reaching their final target,
Table 20. Recommended Starting Dosage Ranges of Intrathecal
Medications for Long-Term Therapy Delivery.
Drug Recommendation of starting dose*
Morphine 0.1–0.5 mg/day
Hydromorphone 0.01–0.15 mg/day
Ziconotide 0.5–1.2 mcg/day (to 2.4 mcg/day per
product labeling)
Fentanyl 25–75 mcg/day
Bupivacaine 0.01–4 mg/day
Clonidine 20–100 mcg/day
Sufentanil 10–20 mcg/day
*Starting doses of continuous intrathecal delivery should be half of
111

Line 6 III I Weak the trial dose for opioid-based medications.

www.neuromodulationjournal.com C 2017 International Neuromodulation Society

V Neuromodulation 2017; 20: 96–132
 DEER ET AL.

histochemical profile, positive epithelial membrane antigen. In the

Table 21. Recommended Doses for Intrathecal Bolus Trialing.
pial layer, collagen fibers and fibroblasts continue under the pial
cells to form the subpial compartment. The trabecular arachnoid
Drug Recommended dose*
surrounds the structures inside the subarachnoid space, including
Morphine 0.1–0.5 mg the spinal cord, nerve roots and vessels that are found free within
Hydromorphone 0.025–0.1 mg the space, providing cover sheaths to these structures (144,145).
Ziconotide 1–5 mcg These sheaths are very fragile and break easily if dissected. The char-
Fentanyl 15–75 mcg acteristics of the arachnoid sheaths in the cauda equina are variable;
Bupivacaine 0.5–2.5 mg some are lax while superimposed planes of the same components
Clonidine 5–20 mcg
with a more compact appearance form others. The thickness of an
Sufentanil 5–20 mcg
arachnoid sheath ranges from 10 to 60 mm (144,145). In some cases,
*Starting doses of medication in the opioid-naive patient for outpa- a single arachnoid sheath encloses one or more nerve roots and in
tient bolus delivery do not exceed 0.15 mg morphine, 0.04 mg hydro- others the nerve root has no sheath at all (144–146).
morphone, or 25 mcg fentanyl.
It is possible that a microcatheter, with small diameter, could be
introduced inside the arachnoid sheath. By contrast, a 20G catheter,
used commonly in epidural techniques, is more difficult to intro-
the substance of the spinal cord. Dura mater, arachnoid and pia duce. If a drug is injected inadvertently or by accident inside the
mater are differentiated structures morphologically, with different sheath, the drug would have a limited dilution with CSF and, there-
properties and, therefore, must be considered and analyzed fore, could potentially be neurotoxic. Taking into account the meth-
independently. od of administering drugs, continuous injection of local anesthetic
Dural Sac through a microcatheter into these arachnoid sheaths could poten-
The dura mater is the most external layer of the dural sac and is tially be more devastating than a single injection. This is because
responsible for 90% of its total thickness. This fibrous structure, aris- repeated doses of small volumes may be accommodated inside the
ing from the meningeal fibroblast, represents a collagenous mem- sheath, leading to nerve damage. The injection of a single larger vol-
brane that confers a barrier to diffusion defined by the molecular ume instead would promote leakage of the anesthetic outside the
weight of the compound crossing the membrane (136). The remain- sheath, decreasing its potential for injury.
ing internal 10% of the dural sac is formed by the arachnoid lamina,
Lumbar Subarachnoid Ligaments
which is a cellular lamina that adds very little extra mechanical resis-
Trabecular arachnoid and subarachnoid ligaments may be related
tance to the compound movement (137). The arachnoid lamina is
to embryonic tissue remnants found in the subarachnoid space,
semipermeable, and influences the passage of lipophilic substances
where the cellular component is progressively replaced by fibrous
through the dural wall (136). The arachnoid limits the diffusion of
connective tissue. These ligaments anchor the lateral, anterior and
injected drugs to the epidural space. Dura mater has a thickness of
posterior sides of the spinal cord to the dural sac (147,148). Sub-
about 0.35 mm (0.25–0.40) (138) that it is fairly constant along the
arachnoid ligaments are similar to trabecular arachnoid, although
length of the spinal cord, with some small variations. It is comprised
contain more collagen fibers, and therefore more resistant to
of concentric dural laminas containing fibers distributed randomly in
mechanical forces. A number of 21 dentate ligaments from each
all directions (139–142). The arachnoid lamina has a thickness of 50–
side of the spinal cord hold to the dural sac. Each ligament is com-
60 microns (mm) (143). Its barrier effect is due to arachnoid cells
posed of a flat fibrous membrane between the anterior and the pos-
strongly bonded by specific membrane junctions. This cell layer rep-
terior nerve roots, and its medial edge is in direct contact with
resents a small thickness of about 10–15 mm.
subpial tissue covering the spinal cord. Laterally, these ligaments
Trabecular Arachnoid give rise to pyramidal projections that attach nonuniformly to the
The trabecular arachnoid originates from the stratum of inner cells arachnoid lamina. The most cephalic ligament is found opposite the
of the arachnoid lamina. These cells surround bundles of collagen margin of the foramen magnum between the vertebral artery and
fibers that form the axis of the arachnoid trabeculae. Near the spinal the hypoglossal nerve. The lowest dentate ligament lies between
cord, the arachnoid cells of the trabecular structure are mixed with the exit of the 12th thoracic and first lumbar spinal nerve roots; this
pial cells from the pia mater. Both types of cells share the same ligament is a thin band stretching downwards from the medullary
cone. Less commonly, posterior ligaments (posticum) are found giv-
ing shape to thin, inconsistent bands that attach the spinal cord to
Table 22. Maximum Concentrations and Daily Doses of Intrathecal the inner surface of the dural sac (147,148). There are also less resis-
Agents as Recommended by PACC 2012 (8) and 2016. tant fenestrated posterior-lateral ligaments, extending more laterally
from the dorsal roots to the arachnoid lamina. Both posterior and
Drug Maximum concentration Maximum dose per day posterior-lateral ligaments extend longitudinally from the cervical to
Morphine 20 mg/mL 15 mg the midthoracic or lumbar level. The thinner ventral ligament is
Hydromorphone 15 mg/mL 10 mg found in the anterior side of the subarachnoid space. These sub-
Fentanyl 10 mg/mL 1000 mcg arachnoid ligaments do not limit free flow of CSF in most patients,
Sufentanil 5 mg/mL 500 mcg due to their discontinuity along the dural sac.
Bupivacaine 30 mg/mL 15–20 mg*
Clonidine 1000 mcg/mL 600 mcg Pia Mater
Ziconotide 100 mcg/mL 19.2 mcg The structure of the pia mater includes a cellular layer and a sub-
pial compartment. The cellular layer is made of flat overlapping pial
*May be exceeded in end-of-life care and complicated cases as deter-
cells with a smooth and bright appearance. It is three to five pial
112

mined by medical necessity.

cells thick (10–15 mm) at medullary level and two to three cells thick

www.neuromodulationjournal.com C 2017 International Neuromodulation Society

V Neuromodulation 2017; 20: 96–132

Table 23. Cerebrospinal Fluid Volume and Nerve Root Volume (mL) per Vertebral Segment (159).
Sacral L5
Cerebrospinal fluid (mean 6 SD) 2.2 6 0.6 4.8 6 1.3
Nerve root (mean 6 SD) 0.1 6 0.1 0.6 6 0.3
*Includes spinal cord volume.
(3–4 mm) at nerve root level. Amorphous fundamental substance is
found around pial cells. The cells measure on average 0.5–1 mm
(149,150).
The subpial compartment has large amounts of collagen fibers,
amorphous fundamental substance, fibroblasts, and a small number
of macrophages, as well as blood vessels. The subpial compartment
is enclosed between the pial cellular layer and a basal membrane in
contact with neuroglial cells. The subpial compartment from low
thoracic vertebrae has a thickness of 130–200 mm, and here mea-
surement variations are more significant than in the pial cellular lay-
er. The thickness of the pia mater is reduced to 80–100 mm at the
level of the medullary cone and continues to diminish down to 50–
60 mm in the origins of the cauda equina. At nerve root level, the
thickness of the subpial compartment is 10–12 mm (149,150).
At the level of the medullary cone, there are perforations or fenes-
trations over the entire surface of the cellular layer of the pia mater.
These fenestrations have circular, ovoid, or elliptical shapes. While
the dimensions of these fenestrations vary, most of them measure
12–15 mm in length and 4–8 mm in width. At nerve root level, the
pia mater also shows similar fenestrations but smaller in size (1–
4 mm) (144,149,150).
Surrounding the pial cells there are numerous macrophages. The
macrophages and other inflammatory cells seen within the pia
mater could possibly originate from subpial and subarachnoid blood
vessels, although a small proportion of them could originate from
immature pial cells as a result of an unknown stimulus. Probably the
fenestrations found in the pia mater are related to the migration of
some immature pial cells as part of an inflammatory response (151).
The number of cell junctions between pial cells is much lower
than among arachnoid cells. For this reason, pia mater is a perme-
able structure allowing the passage of drugs through intercellular
spaces. However, in the area of the conus medullaris the permeabili-
ty could be higher if the fenestrations are present in the patient.
Cerebrospinal Fluid
The volume of the CSF has obvious relevance as a determinant of
dilution of drugs in the subarachnoid space (152). About 500 mL of
CSF is formed each day, mainly by the choroid plexuses of the cerebral
ventricles with uncertain contribution from ependyma, pia, and brain
parenchyma. A small proportion of CSF leaves the skull and enters the
spinal subarachnoid space, passing downwards, posterior to the spinal
cord and returning upwards, anterior to the spinal cord (153), with lit-
tle bulk flow. The rate of absorption through the arachnoid villi varies
and is adjusted to maintain a pressure within normal range.
There are oscillations of the CSF pressure, which are synchronized
with intracranial arterial pulsations for both respiratory and circulato-
ry motors. These changes of pressure could help the dilution of
drugs injected in the CSF to reach a homogenous concentration
around the nerve roots and spinal cord. Their amplitude is about
9 mm per cycle in the cervical CSF and about 4 mm at the thoracic-
lumbar junction, with minimal movement in the distal part of the
lumbar sac (154,155). Pulsations probably increase with the
www.neuromodulationjournal.com C 2017 International Neuromodulation Society
V Neuromodulation 2017; 20: 96–132
INTRATHECAL THERAPY BEST PRACTICES AND GUIDELINES
L4 L3 L2 L1 T12
5.1 6 1.1 4.9 6 0.8 5.7 6 0.9 5.8 6 1.6* 4.76 1.3*
1.3 6 0.3 1.8 6 0.5 2.0 6 0.8 2.4 6 0.5* 2.4 6 0.6*
elevation of intra-abdominal pressure. The oscillatory CSF pulsations
have a significant impact on the spreading of drugs after subarach-
noid injection (156). CSF flow dynamics reveal latencies of the systol-
ic and diastolic peaks of cervical and lumbar CSF pulsations, contrary
to the hypothesis of a continual wave theory. Furthermore, fast flow
velocity reappears in the thoracolumbar spine, correlating to a large
respiratory influence in the thoracolumbar spine (18).
Magnetic resonance imaging (MRI) allows the estimation of CSF
volumes from human axial images under physiological and patho-
logical conditions (157–160). There is a great variability of CSF vol-
ume between patients, although this variation also depends on the
method used to study the CSF. Sullivan et al. (159) in 2006 estimated
a CSF volume of 35.8 6 10.9 mL (range 10.6–61.3 mL) between a
perpendicular plane in the intervertebral midpoint of T12 to L1 and
the lowest limit of the dural sac. Edsbagge et al. (160) in 2011 stud-
ied the complete spine and found a total CSF volume of 81 6 13 mL
(52–103 mL). In the cervical region, there was 19 6 4 mL, in the tho-
racic region, 38 6 8 mL, and in the lumbosacral region 27 6 8 mL.
Another group estimated that the total volume of CSF from L5-S1 to
T11-T12 was 36.1 6 6.7 mL (161). These individual differences of CSF
volume affect the final concentration of a local anesthetic drug
administered in the dural sac of different individuals, even with the
same dose, volume and concentration given. Therefore, having con-
sidered other relevant factors such as position of the patient or ver-
tebral level selected for subarachnoid injection, it may be that doses
below 7.5 mg of bupivacaine do not ensure an adequate level of
blockade in all patients.
Nerve Roots, CSF, and Subarachnoid Catheters
The relationship between CSF volume and nerve root at each ver-
tebral level is an unknown subject that may be of interest when we
consider the concentration of drugs in CSF and the amount of nerve
tissue that must be crossed (161–163). In the cadaver it is possible to
measure the volume of each nerve root, but more difficult to deter-
mine the amount of CSF related to each nerve root. Recently Prats-
Galino et al. estimated the volumes in the segments from L5-S1 to
T11-T12 (162). The total volume of CSF was 29.95 6 5.66 mL
(Table 23) and the volume of the nerve roots was 10.38 6 2.4 mL.
The total mean volume of CSF at each lower thoracic and lumbar
level is around 5 mL per segment, but with a wide range of results
between the different levels (162) (Table 23).
The existence of concentration gradients for many compounds in
the CSF along the spinal canal was established in the 1990s
(139,140). These gradients imply de facto that the CSF cannot be a
circulatory system, and recently MRI evidence has confirmed that
movements of the CSF through oscillations rather than a flow (139).
The velocity of CSF oscillation waves has been calculated at 4.6 m/
sec (SD 1.7 m/sec) (140). This new knowledge has been surprising to
many practitioners and should be considered in all future pharmaco-
kinetic studies of IT drugs.
Very little is known about the mechanisms of drug dispersion in
113
the human CSF. The best available in vivo data are still derived from
 DEER ET AL.
a sophisticated porcine model developed by Bernards (16,20,152).
Overall these studies suggest that all drugs that have been tested
for IT infusion and are relevant to the algorithms are poorly dissipat-
ed in the CSF. This is due to a combination of factors including the
low kinetic energy (low flow rate) provided by all implantable
pumps currently on the market, limited liquid/liquid diffusion coeffi-
cient, and the absence of meaningful CSF bulk flow. Local IT mecha-
nisms related to the interaction of the oscillation of the CSF across
obstacles, such as nerve roots and ligaments, as well as irregularities
in diameters and contours of the walls of the spinal canal, combine
to form regional vortices that increase the dispersion of drugs by
several orders of magnitude (164).
CSF Dynamics
Factors that affect the dynamics of CSF flow include heart rate,
blood pressure (156), and the pulmonary ventilation, which appears
to be one of the most important drivers (165). Traditional knowl-
edge and experience with local anesthetics, opioids, and novel
agents administered in the lumbar spinal subarachnoid space for
surgical anesthesia and analgesia unfortunately contribute little
understanding to the complexities of IT catheter-targeted drug
delivery for chronic pain and spasticity management. In the former,
technical factors such as bolus injection speed or rate, volume, baric-
ity, and patient factors such as height and posture, significantly influ-
ence drug distribution. In the latter, drug distribution is highly
influenced by CSF flow dynamics, where multiple recent studies
have led to a new understanding of drug distribution, which in time
may lead to improved efficacy and safety of IT drug delivery.
The intended action of intrathecally delivered drugs is the spinal
cord and to a lesser extent the brain, and they are delivered into the
CSF through a catheter connected to an implanted pump. The CSF
is secreted by the choroid plexus and brain parenchyma at a rate of
0.3–0.4 mL/min and the total volume ranges from 90 to 150 mL.
Approximately a third of the total volume is contained in the com-
pliant spinal subarachnoid space. Choroid plexus and brain paren-
chyma are not the only sites of CSF production, as glial cells, water
transporters know as aquaporins, and other bidirectional mecha-
nisms produce flow rates greatly exceeding traditional net CSF
secretion rates (166). Absorption traditionally has been considered
to take place at arachnoid villi, and the resultant bulk CSF flow was
thought to influence drug distribution to the spinal cord and brain
to a greater extent than by simple molecular diffusion. This bulk
flow concept has been shown to be outdated from several perspec-
tives, however.
Current understandings from imaged-based and computational
fluid dynamics (CFD) show CSF to behave as a poorly mixed volume
with little net flow, but significant oscillatory flow, originating from
CSF pulsations, which are, in turn, influenced by blood pressure,
stroke volume, and intrathoracic pressure variations associated with
respiration. Because of the noncompliant skull, and to a lesser extent
spinal canal features, these pulse volumes are transmitted as CSF
pulses, and in the compliant spinal canal lead to oscillatory inflow
and outflow at velocities up to 10 mm/sec (167). This pulsatile mix-
ing is orders of magnitude greater that that seen with simple molec-
ular diffusion (164).
This oscillatory flow interaction with the various IT structures such
as nerve roots, ligaments, and objects such as catheters induces sec-
ondary flow patterns known as: 1) steady streaming, which may be
more or less than oscillatory flow, but also greater than molecular
114
diffusion (167), and 2) enhanced diffusion caused by shear forces at
www.neuromodulationjournal.com C 2017 International Neuromodulation Society
V Neuromodulation 2017; 20: 96–132
liquid/solid interfaces (168). Velocity and amplitude variations also
occur in various locations within the spinal canal (169).
In a small-volume (spinal anesthetic) CFD model, speed of drug
transport (i.e., mixing) was strongly affected by the frequency and
volume of CSF pulsations (156). Large-volume bolus injection of
more than 10% of total estimated CSF volume has been shown to
result in rapid substantial mixing throughout the entire spinal axis
independent of pulsatile mixing, but this is orders of magnitude
greater than clinically used in simple continuous or intermittent
patient-activated bolus mode through an implanted catheter (164).
Many animal studies have shown a rostral-caudal gradient from
the catheter tip (20,81,91), and the recent work of Wallace and Yaksh
confirms this in a human study (170). In patients receiving IT mor-
phine, CSF morphine concentrations decreased by distance from the
catheter tip with a gradient that correlated with the infusion dose,
and over a range of infusion rates of 0.1–1.0 mL/day.
Just as volume and flow rate of injection in spinal anesthetics
compared to continuous catheter infusion limit the applicability of
spinal anesthetic data, such discrepancies may also explain the
apparent initial “failures” of IT drug delivery when the catheter loca-
tion, volume, and rate of delivery vary from the trial methodology
and lead to varying and lesser drug distribution reflecting pharma-
cokinetic differences. This has been described with the lipophilic
drug bupivacaine (171), and when simple addition of lipophilic bupi-
vacaine to an existing hydrophilic opioid pump mixture was shown
to be of no benefit (172).
Given all these variables, the amount of drug present at a particu-
lar site along the neuraxis distant from the injection site is difficult to
determine and is not likely to be uniformly distributed from a CFD
perspective (167). Experimentally, drug distribution is limited to a
few centimeters around the tip of the catheter (171,173), and disper-
sion around the cord is also limited (90). This leads to factors to con-
sider when placing a permanent catheter.
Catheter Location and Placement
The traditional teaching of many instructors has reasonably rec-
ommended placing the catheter tip close to the target receptors of
the spinal segment(s) associated with the dermatome/sclerotome/
viscerotome of the primary pain generator. This issue of delivery
location reflects on: 1) the fact that analgesic medications (mu opi-
oid, alpha 2 adrenergic, N type calcium channel blockers), aside
from the local anesthetics, exert their effects on the target recep-
tors/channels that are located on the terminals of the primary affer-
ent and at the level of the first order spinal synapse; and 2) the need
for the drug to reach the spinal levels associated with the spinal seg-
ments processing the pain information (where the target receptors
are located), and the absence of robust infusate redistribution (as
discussed in the previous sections). The ability to determine this
location is, for the most part, difficult, and may be easier to deter-
mine when local anesthetics are administered (174); it is far less clear
when baclofen is infused to treat spasticity or morphine to relieve
pain. An important issue is the appreciation that the receptors asso-
ciated with the target dermatome are not restricted to the spinal
segment associated with the root dermatome. Afferent input into
any given segment may send collaterals up to several segments ros-
trally and caudally. It has been argued that the IT drug must accord-
ingly reach the cells and afferent terminals in these distal
dermatomes (112,175,176).
Consensus Point 23. Limited data exist as to appropriate and
best catheter tip placement. The catheter should ideally be centered

in the spinal dermatome associated with the pain generator. The
consensus recommendation is that the doctor use clinical judgment
based on the clinical setting.
Pharmacokinetics of IT Analgesic Agents
Compared to epidural administration, IT administration has long
been shown to result in higher analgesic efficacy and lower rates of
treatment failures and technical complications (177). The principal
advantage of IT therapies involves bypassing the blood-brain barrier.
This results in higher concentration of administered agents in the
CSF while using lesser amounts of medication. Evidently, greater effi-
cacy is realized with IT drugs that do not freely cross the blood-brain
barrier and when the target receptor is predominantly located in the
CNS in close proximity to the administered IT agent. Medications
deposited in the IT space, through catheters placed near the level
processing the patient’s pain, lie close to but not at the target sites.
Except for local anesthetics and baclofen, the receptor sites for IT
drugs are located in the dorsal horn of the spinal cord, in particular
in lamina II, also known as substantia gelatinosa (178). In order to
reach their target receptors at neuronal synapses in the superficial
dorsal horn, intrathecally administered medications must diffuse
across the pia arachnoid and white matter of the spinal cord—a dis-
tance of up to 1–2 mm from the surface of the cord (179). Diffusion
across the pia is typically considered to be unimpeded given its
structure of a single layer of cells without intercellular junctions.
Work with large molecules has, however, suggested that the diffu-
sion of drugs into the parenchyma may occur though fluid path-
ways paralleling the intraparenchymal vasculature (180). The spinal
cord white matter consists of myelinated axons making it hydropho-
bic (e.g., dorsal column), whereas the grey matter consists of cell
bodies in the various laminae and is hydrophilic (181). Continuous IT
infusion results in stable CSF drug concentrations, which establishes
a gradient driving parenchymal diffusion into the spinal cord. At
equilibrium with slow constant IT infusion, concentrations of small
molecules in the CSF are thought to be equivalent to those in the
interstitial fluid in the superficial aspect of the dorsal horn (179).
A number of factors intrinsic to the IT medication play important
roles in determining drug uptake. Among these, lipid solubility and
molecular weight are the most important physicochemical charac-
teristics of an intrathecally administered drug. Hydrophilic medica-
tions administered intrathecally may have a clinical advantage over
hydrophobic or lipophilic IT agents. Compared to lipophilic medica-
tions, hydrophilic agents have longer half-lives, reflecting the faster
clearance into the vasculature demonstrated by lipophilic agents
(16) and smaller volumes of distribution, resulting in potentially
deeper cord penetration and more rostral spread (182). However,
lipophilic medications have the advantage of limited spread when
precise targeted delivery is desired.
Safety and the Compounding of IT Drugs
Medications approved by the FDA have undergone comprehen-
sive testing in animal and human subjects to demonstrate safety
and efficacy, while the manufacturing process is continuously evalu-
ated to ensure that high quality standards are met (183). Despite
these good manufacturing processes and preclinical testing of medi-
cations before FDA approval, there remains considerable need for
custom formulations of medications that are not commercially avail-
able for IT use or commercially unavailable concentrations of medi-
cations. The practice of creating these mixtures of medications,
known as pharmacy compounding, is not regulated by the FDA but
rather by state boards of pharmacy, incorporating the United States
www.neuromodulationjournal.com C 2017 International Neuromodulation Society
V Neuromodulation 2017; 20: 96–132
INTRATHECAL THERAPY BEST PRACTICES AND GUIDELINES
Pharmacopeia (USP) chapters Pharmaceutical Compounding—non-
sterile and sterile preparations (183). As the FDA does not regulate
these processes, quality assurance is left to the individual pharmacy
or to national compounding associations that offer credentialing.
The FDA defines pharmacy compounding as combining, mixing or
altering of ingredients to create a customized medication for an indi-
vidual patient in response to a licensed practitioner’s prescription
(183). If a physician chooses to use a compounding pharmacy, the
physician should be familiar with quality control procedures of that
pharmacy.
The USP classifies manipulations of sterile products in aseptic con-
ditions as low-risk compounding; however, addition of nonsterile
components would constitute high-risk compounding (183). With
regard to IT drug delivery (IDD), dilution of commercially available
products such as Infumorph (Baxter Health Care, Deerfield, IL, USA)
would constitute low-risk compounding, whereas combining an
aseptic product with a powder formulation, such as bupivacaine,
may constitute high-risk compounding. Beyond the quality assur-
ance issues that lie with the individual compounding pharmacy (and
are largely out of the control of the prescribing practitioner), there
has been considerable discussion about the role of compounded
medications for IDD (8,183,184). It should be noted that drug dilu-
tion is also considered an off-label use of IT medication. This fact,
along with lack of efficacy, has led to the off-label use in the majority
of IT clinical practice.
The PACC of 2012 commented on the role of compounding, iden-
tifying the risks associated with the practice and outlining basic con-
siderations surrounding the practice, such as training of personnel,
segregated sterile compounding facilities, air quality of the com-
pounding area, certification and calibration of equipment, standard-
ized disinfection and quality assurance programs (8). Debate
concerning the practice of using compounded medications contin-
ues to this day, since essentially all medications could be construed
as compounded to some degree (185). Around this time, one IDD
manufacturer reported that off-label medications or admixtures
could result in corrosion to the infusion system and device failure
(186). This bulletin suggested that preservative-free morphine (maxi-
mum approved concentration 25 mg/mL) was approved for use,
however, the bulletin also stated that compounded formulations of
baclofen and morphine had resulted in motor stall, leading to some
confusion on the part of practitioners as to what constituted safe
use of morphine.
In 2013, a joint statement by thought leaders from NANS and the
American Society of Interventional Pain Physicians presented the
viewpoint of many physicians experienced in IDD: that medication
formulations with hydromorphone, fentanyl, and other opioids are
more effective than morphine and have fewer side effects (185). In
addition, this statement reconfirms the use of admixtures of bupiva-
caine and clonidine as outlined in the PACC of 2012 (8). The stall
rate for Synchromed pumps (Medtronic Inc., Minneapolis, MN) was
reported as 2.4% for approved medications (Infumorph, Lioresal,
and Prialt) at five years and as 4.5% for unapproved medications
(185). The PACC suggested that compounded medications were the
de facto standard of care, and peer-reviewed literature exists to sup-
port use of both on- and off-label medications.
An infinite number of drug combinations exist and some physi-
cians recommend that drug mixtures be utilized only where drug
stability information is available. If a study suggests that a high-
concentration drug combination is stable, stability can be assumed
for lower-concentration combinations of the same drugs. For most
pharmaceuticals, there are established and published standards of
115
solubility at room temperature. While high-concentration drug
 DEER ET AL.
combinations do allow for longer refill interval and delivery of higher
daily doses, alterations in pH in high drug-concentration solutions
can lead to pump and catheter failure and patient symptoms. For
example, precipitant has reportedly caused catheter obstruction,
pump corrosion, and failure of IT drug therapy (187,188).
Given that few clinicians utilize IT medications that are completely
devoid of some pharmacy manipulation (dilution, concentration,
etc.) and that most clinicians utilize compounded off-label medica-
tions and admixtures, IDD essentially mandates use the of com-
pounded medications. Also, since the incident rate of motor stalls
for approved and nonapproved medications is similar, it seems pru-
dent that diligence in monitoring patients receiving IDD for the pres-
ence of motor stalls or therapy disruption be underscored to detect
and treat these possible outcomes.
Review of IT Medications
To date there has been considerable controversy and little consis-
tency in the trialing of opioids, although trialing with morphine or
hydromorphone is common and advised (27,59). The use of pretrial-
ing systemic opioid dose conversions to derive an appropriate dose
for IT opioid trialing is not recommended because of pretrial wean-
ing of systemic opioids and differences in pharmacology between
systemic and IT opioids.
Morphine
Mechanism of Action. Morphine is the most widely used IT medi-
cation. It is a mu opioid agonist (189).
Neurotoxicity. Preclinical evaluation in several large animals mod-
els showed morphine’s propensity to initiate space occupying
masses or IT granulomas (190–193). These masses, constituted of
fibroblasts, maturing collagen and interspersed with inflammatory
cells, arise from the dura/arachnoid, with the mass size largest proxi-
mal to the catheter tip (188,194). This profile has been observed
with several opiates, including hydromorphone and methadone
(131). The association between IT opioid therapy and granulomas is
further discussed in a PACC 2012 publication (60).
Clinical Studies. Clinical data on IT morphine continue to support
its use as a first-line therapy. From 1983 to 2000, there were many
studies showing efficacy of the long-term infusion of morphine and
morphine/adjuvant admixtures, as reported by PACC 2012 (8).
Recent results from several long-term studies support the efficacy of
IT morphine in treating patients with chronic pain, including pain
from cancer and noncancer diseases. In a retrospective study, medi-
cal records from 57 patients with chronic malignant pain on long-
term IT opioid therapy (morphine, hydromorphone, or sufentanil)
were reviewed (195). VAS scores for pain significantly decreased
from baseline to time of first refill (p  0.001); VAS scores then
remained stable and significantly lower than baseline scores
(p  0.001) through year 3. Oral opioid use decreased significantly in
the first year of IT therapy (p  0.001) and increased slightly but
insignificantly between years 1 and 3.
In a prospective, open-label study of IT morphine infusion (Prom-
etraVR Infusion Pump, Flowonix Inc., Mt. Olive, NJ, USA), 110 patients
with chronic pain were treated and followed up for approximately
one year (196). Pain relief was noted within one month and was sus-
tained during the following six months; trends indicated consistent
pain relief through 12 months. In an open-label study, 13 patients
with intractable pain from chronic pancreatitis who had undergone
a successful trial of IT opioids received IT opioid infusions for a mean
116
duration of 29 months (197). The limited intention-to-treat analysis
www.neuromodulationjournal.com C 2017 International Neuromodulation Society
V Neuromodulation 2017; 20: 96–132
revealed an overall success rate of IT opioid therapy of 76.9% of
patients. In another open-label study, IT morphine was infused in 24
patients with vertebral fractures due to osteoporosis who had not
responded to systemic opioid therapy (198,199). The mean VAS pain
score decreased significantly from 8.7 cm before IT therapy to
1.9 cm after one year of IT therapy (p < 0.001). Significant improve-
ments from baseline to one year were also noted on scores for the
Quality of Life Questionnaire of the European Foundation for Osteo-
porosis subscales for pain, quality of daily life, domestic work, ambu-
lation, and perception of health status (p < 0.001).
In one retrospective study, investigators attempted to determine
characteristics of patients for whom IT morphine therapy is effective
(200). The study included 131 patients who received IT morphine
monotherapy for various pain types (cancer-related, nociceptive, or
neuropathic). A >50% decrease in pain was reported in 73% of all
patients. No differences in responder rates were noted when results
were analyzed by pain type, patient age, or morphine dosage; how-
ever, responder rates were significantly higher in men than in wom-
en (p 5 0.02).
Raphael et al. conducted a randomized, double-blind controlled
trial of IT morphine efficacy in noncancer pain (201). One group had
no change in morphine dose while the other had a 20% reduction
every week for ten weeks. Seven of ten patients, all in the dose-
reduction group, withdrew from the study prematurely. Within-
group VAS and Oswestry Disability Index differences were statistical-
ly significant between baseline and the last observation for the inter-
vention group, with statistically significant greater pain and
worsened disability in the dose-reduction group. These results sug-
gest the efficacy of IT morphine for long-term treatment of non-
cancer pain.
Hydromorphone
Mechanism of Action. Hydromorphone is a mu opioid agonist
(202).
Neurotoxicity. Preclinical studies of IT infusion of hydromorphone
in large animal models showed space occupying granulomas at
higher concentrations (131,203).
Clinical Studies. The literature review revealed no new studies
investigating the efficacy of IT hydromorphone in the treatment of
chronic pain. Two case reports described granuloma development
in patients treated with IT hydromorphone. The first described a
patient who developed a granuloma on IT morphine; nine months
after removal of the first granuloma, she developed another granu-
loma after one month of IT hydromorphone therapy (204). The sec-
ond report described a 52-year-old man with a history of chronic
lumbar spine pain who developed a granuloma while receiving
high-concentration IT hydromorphone (85 mg/mL) at a dose of
19.8 mg/d (205).
Mallinkrodt Phamraceuticals (St. Louis, MO, USA) is enrolling
patients in a study to develop a branded and formally manufac-
tured, FDA-approved hydromorphone product. The first trial is a
controlled, two-arm, parallel-group, randomized withdrawal study.
Subjects in this trial will already have implanted IT pumps and will
be transitioned to IT hydromorphone. They will then be titrated to a
level where oral opiate medications are eliminated up to a dose of
5 mg of IT hydromorphone per day. Subjects who attain stabilization
and meet criteria for randomization will be assigned to either remain
at their current dose of hydromorphone or be titrated off therapy in
a blinded fashion. The primary efficacy end point of this study is the
proportion of subjects who are treatment failures during the

double-blind randomized withdrawal period. The second trial is a
Phase 3, open-label, single-arm safety study where subjects can
either directly enter the study from the previously described trial or
be transitioned directly to IT hydromorphone. The two formulations
of IT hydromorphone being investigated are a 2 mg per cc and a
10 mg per cc concentration, respectively. These trials are actively
recruiting subjects and results are expected in 2017.
Peripheral edema associated with IT hydromorphone infusion was
reported in one patient. This 61-year-old woman with chronic pain
developed progressive lower extremity edema, which was compli-
cated by severe cellulitis, while on IT morphine (206). Her edema
lessened when she was switched to IT hydromorphone but recurred
with severe cellulitis two months later. Her IT regimen was changed
to clonidine (33 mcg/d) and baclofen (67 mcg/d); edema resolved
and did not recur.
Fentanyl
Mechanism of Action. Fentanyl is a lipophilic mu opioid agonist
(132,207).
Neurotoxicity. Preclinical studies of IT infusion of fentanyl or alfen-
tanil in large animal models showed no space occupying granulo-
mas at the highest concentrations examined (132,133).
Clinical Studies. The literature review revealed no new studies
investigating the efficacy of IT fentanyl in the treatment of chronic
pain. One case report described a 34-year-old woman receiving IT
combination therapy (fentanyl, bupivacaine, and clonidine) for
chronic pain who was suspected of having an epidural hematoma
because of inadequate pain control despite increasing doses and an
unsuccessful epidural steroid injection (208). On operation, a
catheter-tip mass was noted within the epidural space, with the
catheter tip in its center. Misplacement of the catheter (epidurally
instead of intrathecally) at the time of original insertion complicated
diagnosis. Histopathological analysis revealed a proteinaceous mass,
which the authors determined was an inflammatory mass, not a
granuloma, and was likely a result of drug precipitation.
Sufentanil
Mechanism of Action. Sufentanil is a potent mu opioid agonist
(209).
Neurotoxicity. No canine continuous infusion trials with sufentanil
have been reported. Repeated bolus delivery in the canine mode
showed no histopathologic changes (132).
Clinical Studies. The literature review revealed no new studies
investigating the efficacy of IT sufentanil in the treatment of chronic
pain. One case report described an 86-year-old woman with FBSS
who had received multiple IT therapies over the course of two years
(210). Six weeks after beginning IT sufentanil therapy (12–17.2 mcg/
d), she had lower extremity weakness, sensory changes, and intracta-
ble lumbar pain, and a CT-myelogram demonstrated the presence
of a granuloma. Sufentanil was removed from the pump and
replaced with normal saline. Her symptoms resolved within approxi-
mately 48 hours, and the patient was receiving oral methadone ther-
apy for pain at the time of hospital discharge.
Ketamine
Mechanism of Action. Ketamine is a noncompetitive antagonist
that blocks the glutamate NMDA ionophore (211).
www.neuromodulationjournal.com C 2017 International Neuromodulation Society
V Neuromodulation 2017; 20: 96–132
INTRATHECAL THERAPY BEST PRACTICES AND GUIDELINES
Neurotoxicity. IT ketamine infusion (10 mg/mL delivered at 2.4 mL/
d) in chronically catheterized dogs resulted in mild to severe spinal
pathology ranging from local demyelination to necrotizing lesions
of spinal parenchyma near the catheter tip (10). Similar pathology
was observed in neonatal rats (213). This effect was shared by other
N-methyl-D-aspartate (NMDA) antagonists, including MK801, mem-
antine, amitriptyline, and S-methadone. Notably, these studies
unfortunately did not establish a no effect level nor correlate the
lower doses with an antihyperpathic action.
Clinical Studies. A randomized, double-blind study compared the
use of epidural ketamine plus bupivacaine vs. epidural bupivacaine
plus saline in 53 patients undergoing lower limb amputation (214).
In both treatment groups, persistent phantom and stump pain were
less than that seen in comparable studies and did not differ signifi-
cantly between groups. In the ketamine/bupivacaine group, signifi-
cant decreases from preoperative anxiety and depression levels
were noted and persisted through the one-year follow-up point.
Additionally, a case report described a 49-year-old woman with
severe cancer-related upper back and abdominal pain (215). Her
numeric rating scale (NRS) pain score was 6, despite 96 days of IT
therapy with a combination of morphine and bupivacaine. IT keta-
mine was added to her regimen and her NRS score decreased to 3.
There were no signs of motor paralysis, psychomimetic alteration,
neurological dysfunction, or infection in this terminally ill patient.
This contrasts with subpial vacuolar myelopathy, which was found
postmortem in a cancer patient treated with 5 mg/day IT ketamine
for three weeks (216).
Methadone
Mechanism of Action. Methadone is a racemic compound in which
the d-isomer has NMDA receptor antagonist activity and the l-
isomer is a mu opioid agonist (189).
Neurotoxicity. Notably, there is concern about the safety of IT
methadone, since all compounds with NMDA activity have serious
neurotoxic effects (212). Continuous infusion of the isomers in a dog
model revealed spinal toxicity and granulomas with either isomer
(131).
Clinical Studies. The efficacy of epidural methadone was investi-
gated in a study of 32 patients with cancer-related pain that was
refractory to epidural morphine (217). Patients received one of the
following treatments: 2.5, 5, or 7.5 mg epidural methadone diluted
in 60 mg lidocaine or 7.5 mg epidural methadone diluted in 60 mg
lidocaine plus 10 mg dexamethasone. Epidural methadone provided
dose-dependent analgesic effects, and these effects were further
improved with the addition of dexamethasone. A prospective study
of IT methadone was performed in 24 patients (218). Thirteen
patients experienced improvement of their pain control with metha-
done, nine continued to receive this agent for six months with good
pain relief.
Ziconotide
Mechanism of Action. IT ziconotide is first-line therapy for both
neuropathic and nociceptive pain, and is FDA approved. Its mecha-
nism of spinal action is to block presynaptic N-type calcium channels
in the dorsal horn of the spinal cord (219,220). This targeting is dis-
tinctly different from mu agonism and allows ziconotide to be help-
117
ful in the opioid-tolerant patient (32).
 DEER ET AL.
Table 24. Recommendations Regarding Intrathecal Clonidine Treatment by the PACC Using USPSTF Criteria.
Statements
Intrathecal clonidine in CRPS patients decreases pain scores over time as
well as allodynia, hyperalgesia, and mean arterial blood pressure.
Clonidine increases analgesia duration and decreases morphine use in the
acute postoperative setting.
Clonidine may precipitate hypotension in patients with baseline
hypertension.
Ziconotide concentration decreases over time when mixed with clonidine.
Neurotoxicity. Ziconotide (Prialt) has undergone extensive preclini-
cal safety evaluation in multiple species (221) without spinal toxicity
in the concentrations employed.
Clinical Studies. Many methods have been employed to trial zico-
notide, from continuous infusion to single-shot bolus. There is no
evidence to suggest one trialing strategy is better than another,
appreciating that on-label FDA trialing is performed by a microam-
bulatory delivery device, or via the Medtronic Synchromed II device.
In patients with neuropathic pain who are not responsive to pre-
trialing systemic opioid therapy, a trial of ziconotide should be con-
sidered. Trialing with ziconotide can be challenging because the
drug’s narrow therapeutic window and the side effect profile is
more closely related to the rate of dosage increase. Thus, trialing
with an externalized catheter may be impractical and hazardous
because of the slow titration required with ziconotide (e.g., dosage
increases of 0.5–1.0 mcg every several days). Although trialing with
bolus dosing can be useful to identify some appropriate candidates,
side effects associated with bolus dosing may eliminate many
patients who might otherwise have benefited from IT ziconotide
therapy. Thus, the use of alternative trialing methods in order to
avoid a trial failure because of intolerable side effects would be
advantageous in this regard. Although meclizine treatment is some-
times used before IT ziconotide trials in clinical practice, there is
insufficient evidence to support this approach. Proper hydration via
intravenous (IV) infusion before trialing may limit the side effect of
hypotension.
Bupivacaine
Mechanism of Action. Bupivacaine is an amide local anesthetic
with high lipid solubility that is often used off-label in IT therapy
(99,134,222,223). Several mechanistic properties give it utility with
spinal delivery: 1) differential efficacy at low concentrations that
alters sensory processing while sparing motor function (224–227); 2)
absence of tachyphylaxis in patients with neuropathic or somatic
pain (174,228–230); and 3) potent synergy with other IT analgesic
targets in animal models (110,231) and in humans (226,227,229,230).
Neurotoxicity. In early rat studies with continuous bupivacaine
infusion, modest increases in neuronal vacuolation, was observed at
concentrations of 0.5% (232). Bolus delivery in dogs of bupivacaine
(0.75%) resulted in minor leptomeningeal cellular infiltration (233). IT
infusion of 2.5–3.8 mL/d of 0.25% bupivacaine for 3–11 weeks
resulted in mild leptomeningeal cellular infiltration in two of eight
animals (233). In combination with morphine in human cancer
patients no significant histopathology was noted on autopsy
(234,235). IT infusion of bupivacaine has a long track record of safety
118
and efficacy alone or in combination with morphine.
www.neuromodulationjournal.com C 2017 International Neuromodulation Society
V Neuromodulation 2017; 20: 96–132
Evidence levels Recommendation strength Consensus strength
I A Strong
II-2 B Strong
II-3 B Strong
II-3 B Strong
Clinical Studies. Though not FDA-approved for continuous IT use,
bupivacaine is the most common local anesthetic used in spinal
anesthesia and is used off-label in IT therapy. Compared to the epi-
dural route, IT drug delivery results in higher patient satisfaction,
fewer catheter complications, better pain relief and sleep (174).
Combinations of bupivacaine and opioids have shown synergistic
efficacy in acute postoperative and labor pain studies (110,236–239).
In chronic pain settings with continuous IT drug delivery, howev-
er, the effect is less clear. A retrospective study of 109 patients, who
were managed with a solution containing a mixture of IT bupiva-
caine and opioids after an initial period of IT opioid-only treatment,
found improved pain control and decreased oral opioid consump-
tion with combination therapy compared to opioids alone (134).
The average daily bupivacaine dose in that study was 10 mg. A small
double-blind randomized prospective study suggested that addition
of bupivacaine (up to 8 mg/day) to IT pumps infusing morphine or
hydromorphone did not result in improved pain control in patients
with low back pain, mostly in the setting of postlaminectomy syn-
drome (172). However, a large study in noncancer pain patients
revealed blunting of opioid dose escalation in IT-therapy patients
receiving bupivacaine in addition to IT opioids. The average bupiva-
caine daily dose in those patients was 9.8 mg at one year postim-
plant (99). Nevertheless, the difference in pain scores between the
group receiving IT opioids and the group receiving IT opioids plus
bupivacaine was not statistically significant. A similar effect of blunt-
ing IT morphine dose escalation was noted in a small cancer pain
study (240).
The high lipid solubility of bupivacaine limits its spread intrathe-
cally and highlights the need to place the catheter in the posterior
IT space at the site of processing-pain pathology (20). No formal
studies have been performed to assess starting and maximal doses
of bupivacaine in IT therapy. In addition, there are no prospective
studies of chronic IT bupivacaine administration as a sole agent. The
most recent PACC guidelines have suggested a maximal concentra-
tion of 30 mg/mL, a starting dose of 1–4 mg/day and a maximal dai-
ly dose of 10 mg (8). However, this maximal dose is similar to the
average dose noted to be effective in previous studies (99,134).
Additional bupivacaine is sometimes self-administered in boluses by
patients through a personal therapy manager (PTM) device (241).
Serious cardiotoxic side effects can occur when significant amounts
of bupivacaine reach the bloodstream. This should not be of con-
cern with IT bupivacaine infusion (224). Clearly, the limiting factor in
bupivacaine infusions would be sensorimotor loss. Nevertheless, IT
bupivacaine doses as high 5 mg/hour (or 120 mg/day) and bolus
doses as high as 7.0 mg in the high cervical IT space have been
reported with no apparent untoward manifestations (242). Catheter
tip location, CSF dynamics, and patient mobility likely play important
roles in sensorimotor loss in response to bupivacaine. The most
recent version of PACC guidelines suggested that, in neuropathic

Table 25. Recommendations Regarding Intrathecal Baclofen Treatment by the PACC Using USPSTF Criteria.
Statement
Baclofen should be considered an intrathecal medication for use to treat spasticity.
Baclofen can be used as an adjuvant to treat pain.
Care regarding mitigating withdrawal from baclofen is suggested.
Ancillary resources regarding physical therapy to aid in titration and assessment when
employing baclofen is recommended.
Using bolus or flex dosing strategies to improve spasticity demonstrates promise.
pain, bupivacaine in combination with morphine is considered first
line but second line in combination with morphine, fentanyl or
hydromorphone in nociceptive pain (8). It should be noted that
there is no basis for such an assertion. Recent data suggest efficacy
of IT bupivacaine in combination with hydromorphone as first-line
therapy (135). Average daily dose of bupivacaine at 24-month post-
implant was 12.1 6 0.9 mg including on average 3.7 6 0.6 PTM
boluses of 0.78 6 0.05 mg bupivacaine each.
Clonidine
Mechanism of Action. Clonidine is an alpha2 adrenergic agonist
(243). Clonidine may exert antiallodynic effects by inhibiting the acti-
vation of glial cells and by activation of nuclear factor jB and p38
(MAP kinase), thus inhibiting the production of proinflammatory
cytokines (244). Increasing evidence suggests that activated spinal
cord glial cells contribute to enhanced pain states through the
release of proinflammatory cytokines, such as tumor necrosis factor-
a (TNF a), IL-1, and IL-6 (245,246).
Neurotoxicity. Results of studies in large animals treated with epi-
dural clonidine for 28 days (concentrations up to 2 mg/mL and
doses up to 7.7 mg/d) revealed no notable histopathologic findings
(245). Additionally, IT infusion of clonidine (2 mg/mL at 2.4 mL/d)
monotherapy for 28 days was not associated with direct evidence of
spinal histopathology (191). In the same study, in dogs that received
admixtures of clonidine and morphine, the severity of spinal histopa-
thology decreased in a clonidine dose-dependent manner.
Clinical Studies. Combination therapy including IT clonidine was
described in a case report of a 79-year-old man with chronic lower
extremity pain (246). Approximately one year after beginning IT
therapy with fentanyl, bupivacaine, and clonidine, the patient
reported night terrors, insomnia, severe dry mouth, and increased
depression. Three days after discontinuation of clonidine therapy,
his depression improved and the other symptoms resolved; the
symptoms have not recurred after >2 years of clonidine-free IT ther-
apy. Clonidine has been evaluated in many studies, with improve-
ment in analgesia and opioid-mitigating effects (247–250).
The PACC recommendations for IT clonidine appear in Table 24.
Baclofen
Baclofen is commonly used for intractable spasticity and is FDA
approved for use in IT pumps. Baclofen has limited use as a mono-
therapy option for the primary treatment of chronic pain. It is used
most commonly in combination therapy to treat pain with spasticity
(Table 25).
Mechanism of Action. Baclofen is an agonist of the gamma-
aminobutyric acid (GABA)-A receptor. In preclinical studies, the
GABA-A receptor, a chloride ionophore, has been shown to exert
www.neuromodulationjournal.com C 2017 International Neuromodulation Society
V Neuromodulation 2017; 20: 96–132
INTRATHECAL THERAPY BEST PRACTICES AND GUIDELINES
Evidence level Recommendation grade Consensus strength
II-2 A Strong
II-3 B Moderate
II-2 A Strong
III C Moderate
II-3 B Moderate
antihyperalgesic effects at the spinal level (251,252). Concurrent
with these effects, baclofen at the GABA-A receptor can have promi-
nent effects on motor tone via direct hyperpolarization of the motor
horn cells.
Neurotoxicity. IT baclofen infusion (at rates of up to 2 mg/mL/d)
for 28 days in chronically catheterized dogs has been shown to
result in no behavioral or spinal histological evidence of neurotoxici-
ty (253). Additionally, preclinical evaluation suggested that IT baclo-
fen at doses up to 2 mg/mL/d were not associated with granulomas
in dogs (254). The development of granulomas in humans is rare
with IT baclofen therapy (255). Granuloma formation was reported
in two patients receiving IT baclofen monotherapy (256). However,
these reports, plus another, were later re-evaluated, and other scien-
tifically plausible explanations (e.g., baclofen precipitation) were
posited for MRI findings in these patients who were originally
reported to have IT baclofen-induced granulomas (257). The associa-
tion between IT baclofen therapy and granulomas is further dis-
cussed in the brief report titled, “Polyanalgesic Consensus
Conference 2 2012: Consensus on the Diagnosis, Detection, and
Treatment of Catheter-Tip Inflammatory Masses (Granulomas)” (60).
Clinical Studies
Neuropathic Pain. Recent reports of the use of IT baclofen for the
treatment of patients with neuropathic pain include two studies and
two case reports. In a double-blind study, the effect of different IT
baclofen infusion rates (i.e., 0.75 or 3 mg/mL baclofen solution
infused at a consistent rate) on pain and dystonia was investigated
in 14 patients with complex regional pain syndrome (CRPS) who
had not responded adequately to previous IT baclofen therapy
(258). Overall, the faster baclofen infusion rate was not associated
with improvements in dystonia or pain but was associated with
increased frequency of adverse events (AEs). However, in a subset of
six patients for whom AEs had previously prohibited dose escalation
of IT baclofen, all but one preferred the faster infusion rate, reporting
that the effects of the faster-infusion IT baclofen on pain and dysto-
nia outweighed the severity of AEs. One report described two cases
of baclofen and ziconotide combination therapy (259). The first
patient was a 48-year-old man with neuropathic pain who had
received ziconotide (2.4 mcg/d) for approximately three months
before baclofen (110–115 mcg/d) was added to his IT regimen for
spasticity control. His ziconotide dosage was then reduced to 1.7
mcg/d over the course of one month. After eight months of zicono-
tide/baclofen therapy, his VASPI score had decreased by 75%. The
second patient was a 73-year-old man with neuropathic pain who
had received ziconotide monotherapy (dosage at onset of pain
relief, 14.4 mcg/d) for six months when baclofen (62 mcg/d) was
added for control of spasticity. After two years on ziconotide/baclo-
119
fen therapy, his VASPI score had improved from baseline by 30%. He
 DEER ET AL.
Table 26. Intrathecal Drug Delivery Systems.
Codman 3000 Medtronic Isomed
Catheter material Polyurethane with titanium Radiopaque silicone rubber
reinforced coil with titanium tip
Pump material Titanium/silicone rubber Titanium/silicone rubber
Pump mechanics Continuous flow propellant Continuous flow propellant
MRI compatibility No effects 3T No effects 1.5T
Patient-controlled intrathecal None None
analgesia (PCITA)
also experienced improvements in mood and ability to perform
activities of daily living during this time.
Trialing. In a study that included 48 patients with neuropathic pain
who had inadequate response to SCS, participants were given IT bac-
lofen boluses (25–100 mcg) (260). Among these patients, 14 were
classified as responders (>50% improvement from baseline in pain
level), and 11 had pumps implanted for continuous IT baclofen infu-
sion (four with pumps alone, seven with SCS plus pumps). Follow-up
after an average of 32 and 67 months of SCS plus baclofen therapy
revealed that >50% of patients maintained good treatment effects;
baclofen doses approximately doubled during this time.
Tolerance. Tolerance/tachyphylaxis is an important consideration
when using IT baclofen, as it may occur in approximately 22% of
patients treated with long-term IT baclofen (261). Tolerance may
develop even after very long-term treatment, as was described in a
case report of a patient who developed tolerance 16 years after initi-
ation of IT baclofen therapy (262). A drug holiday of 24 hours
(with careful monitoring for withdrawal symptoms) may be helpful.
Additionally, limited data in four patients suggest switching to a pul-
satile bolus infusion may help address tolerance (261).
Withdrawal. Abrupt cessation of IT baclofen therapy could result in
baclofen withdrawal, a serious, life-threatening situation that can be
severe and prolonged (263). Baclofen withdrawal may mimic seroto-
nin syndrome (264) and has rarely been associated with hallucina-
tions (265). One case report described baclofen withdrawal after
removal of an IT baclofen pump in a 45-year-old woman with para-
plegia and severe lower extremity spasticity (266). She was treated
with oral baclofen, lorazepam, phenytoin, and tizanidine and gradual-
ly improved over the course of seven days. She was discharged on
phenytoin, linezolid, and metoprolol, with no need for oral spasticity
therapy. It is also important to note that IT baclofen withdrawal may
result from catheter leakage (267). Clinicians should be aware of the
signs and symptoms of baclofen withdrawal and be watchful for
them in any patient who receives IT baclofen. One report described
the successful weaning of a patient from high-dose IT baclofen thera-
py through use of a lumbar drain and standard PCA pump delivering
continuous infusion of IT baclofen as a means of avoiding withdrawal
(268). It should also be noted that symptoms of baclofen withdrawal
might be the first indication of IT catheter migration. Since baclofen
is a water-soluble agent, migration of the catheter into the epidural
space will result in symptoms of baclofen withdrawal.
Overdose. Baclofen overdose is a potentially life-threatening condi-
120
tion, the signs and symptoms of which may include somnolence,
www.neuromodulationjournal.com C 2017 International Neuromodulation Society
V Neuromodulation 2017; 20: 96–132
Flowonix Prometra II Medtronic SynchroMed II
Radiopaque silicone rubber Radiopaque silicone rubber
with tungsten tip with titanium tip
Titanium/silicone rubber Titanium/silicone rubber
Valve gated programmable Peristaltic titanium/
plastic programmable
MRI conditional 1.5T MRI conditional 3T
with valve shut-off
Patient therapy Personal therapy
controller (PTC) manager (PTM)
hypotonia, seizures, autonomic instability, bradycardia, and respira-
tory depression (269). One case report described baclofen overdose
associated with a change in IT baclofen concentration combined
with the performance of a catheter dye study (269).
Combinations of IT Drugs
A number of studies have been conducted to evaluate the use of
IT morphine in combination with other IT agents, such as bupiva-
caine, ziconotide, and baclofen. One such open-label study included
55 patients with advanced cancer-related pain who had been unre-
sponsive to previous trials of systemic opioids alone and were
treated with a combination of IT morphine and IT bupivacaine and
followed for up to six months (270). The initial IT morphine dosage
was calculated from the patients’ previous systemic opioid dosage
by using an oral:IT ratio of 100:1 (which is notably different from the
300:1 ratio that is typically used for equianalgesic calculations) (271).
The initial bupivacaine dosage of 12.5 mg/d was increased to
25 mg/d before the IT morphine dosage was increased and modi-
fied as needed. Significant reductions in pain intensity, along with
significant decreases in the mean systemic opioid dose, were noted
at one and three months after initiation of IT therapy and up to the
time of death (p  0.029). In another open-label study, which includ-
ed 32 patients with chronic noncancer pain who had >70% pain
relief after a trial of low-dose IT morphine and bupivacaine, continu-
ous IT therapy (0.1 mg/d morphine, 0.5 mg/d bupivacaine) was initi-
ated, and dosages were titrated to a mean of 1.03 mg/d morphine
and 1.15 mg/d bupivacaine (272). Mean VAS pain scores decreased
significantly from baseline to month 3 (p < 0.01) and remained con-
sistently reduced through the 48-month follow-up.
The addition of IT morphine in 25 patients with suboptimal pain
relief on stable dosages of IT ziconotide was investigated in an
open-label study (273). VASPI scores for these patients improved by
a mean of 26.3% by week 4 of combination therapy, and mean sys-
temic opioid consumption decreased by 49.1%. Notably, stability
data regarding ziconotide and opioid admixtures may provide guid-
ance for frequency of pump refills (115,274).
INTRATHECAL DRUG DELIVERY SYSTEM
CHARACTERISTICS AFFECTING IT THERAPY
Pump and Catheter Materials and Mechanics
Intrathecal pumps can be mainly differentiated into systems that
are continuous flow or variable flow. The driving mechanisms may
include peristalsis, fluorocarbon propellant, osmotic pressure, piezo-
electric disk benders, or the combination of osmotic pressure with
an oscillating piston (Table 26). Pump materials are similar with the
pump shell being titanium and filling ports containing silicone

rubber. Physical orientation of the filling and side ports are largely
consistent, with differences in negative pressure or positive pressure
confirmation strategies (275).
Pump delivery mechanics include continuous flow propellant or
programmable features. Propellant pumps (Codman 3000 and Med-
tronic Isomed) do not require batteries and deliver a continuous
flow for the life of the pump. The programmable pumps require bat-
tery replacement, based on labeling, at a maximum of five to seven
years for the Medtronic Synchromed II and maximum of 10 years for
the Flowonix Prometra II.
The programmable pump systems feature differences that deserve
mention. The Medtronic Synchromed II (Minneapolis, MN, USA) sys-
tem uses a peristaltic rotor system of internal tubing to deliver medi-
cation from the reservoir to the external catheter system. The
Prometra II Flowonix Pump (Mount Olive, NJ, USA) employs a valve-
gated bellow delivery mechanism. Each pump has the ability for
patients to deliver patient-controlled dosing by using a patient-held
programmer (Patient Therapy Manager or PTM for Medtronic and
the Patient Therapy Controller or PTC by Flowonix). Both pumps sup-
port MRI conditional labeling, with the Medtronic pump up to 3 Tes-
la and the Prometra pump at 1.5 Tesla. Of note, both pumps require
interrogation following a scan. For the Medtronic system, exposure
to a magnetic field will create a motor stall, which typically resolves
following removal of the magnet and can occur within 20 min to 2
hours, with a failure of motor stall recovery on rare occasions. For
this reason, it is suggested to interrogate after the scan (276).
The Flowonix Prometra I system requires removal of all medica-
tion within the reservoir prior to MRI exposure, as failure can result
in emptying of the reservoir contents into the patient. The Prometra
II system remedied this concern with the flow activated valve (FAV)
that is triggered when exposed to a magnetic field, blocking drug
delivery from the reservoir to the patient after delivery of less than
or equal to 10 lL (275). If the contents of the reservoir are expected
to be less than 1 mL, they should be removed prior to the MRI
because the FAV may not activate. After the MRI, the contents of the
reservoir have to be removed entirety to manually reset the FAV,
the pump interrogated and the contents replaced in sterile fashion,
with elapsed time of 3 min (275).
The Medtronic Synchromed II System has a minimal flow rate of
0.048 mL/day to allow for programming, while the Prometra II sys-
tem can be at zero flow. Accuracy of the Prometra system is greater
(97.8%) compared to the Medtronic Synchromed II system (2 vs.
14.5%) (276–279). The Medtronic system has two reservoir sizes, 20
and 40 mL, while the Prometra II system has a 20 mL reservoir only.
Although it is beyond the scope of the PACC, the consensus
group felt it necessary to comment briefly on the warning letters
surrounding the Medtronic Synchromed II system, and the recent
consent decree agreement between the FDA and Medtronic in April
2015. Prior to this, warning letters were released regarding overinfu-
sion, corrosion of the internal tubing with the use of off-label medi-
cations or combinations or medications, and priming bolus errors
(280). The complexity this introduces for use of the Medtronic Syn-
chromed II system is unknown, with a recent editorial offering a
foundation for discussion (5).
Intrathecal Infusion Rate
Intrathecal therapy offers advantages over systemic therapy in
that IT delivery bypasses the blood-brain barrier with direct access
of the drug delivered to receptor sites in the dorsal horn of the spi-
nal cord (283). The efficacy of this therapy is dependent to some
degree on drug distribution within the spinal canal; however, the
www.neuromodulationjournal.com C 2017 International Neuromodulation Society
V Neuromodulation 2017; 20: 96–132
INTRATHECAL THERAPY BEST PRACTICES AND GUIDELINES
biophysiologic properties that determine drug distribution in the
spinal canal are incompletely understood. Many factors have been
proposed and evaluated as contributing to differential drug distribu-
tion in the CSF. For instance, anatomic variation, postural changes,
drug solution density, binding characteristics of drugs at the dorsal
horn, CSF volume and variations in CSF pulsatile flow with heart
rate, stroke volume, and respiratory cycle have been examined
(20,156,170,281,283). Additionally, it has been suggested that the
rate of dispersion in the CSF cannot be explained by diffusion alone
(156,283).
CSF convective transport within the spinal canal has been studied
exclusively, and it is now known that CSF flow is pulsatile with oscil-
latory displacements creating microenvironments with eddy cur-
rents resulting in complex micromixing of infused drugs with no net
bulk flow (20,156,281,283). A recent study based on computer
modeling of microanatomic structures in the spinal canal suggested
that the spinal nerve roots themselves serve as a significant barrier
to laminar flow and may create much of the geometric-induced
flow patterns observed and postulated through various experimen-
tal designs (281). With regard to rostral-caudal spread of hydrophilic
medications such as morphine, it has been demonstrated that a
rather steep concentration gradient exists as samples are taken at
points further removed from the catheter tip. The authors of this
finding suggested that drug dilution over distance and drug concen-
tration at the site of action may be important in providing analgesic
efficacy (170). Given these data, simplistic views of CSF dynamics do
not provide insight into the possible clinical implications of decisions
made surrounding drug-infusion rate and drug dispersal within the
CSF. Much of what is known clinically is derived from observations
utilizing spinal anesthesia for surgical intervention (282). Despite this
complexity there are some basic observations that have been
reported in the literature that provide insight into the clinical utility
of IT drug delivery flow-rate manipulations on treatment outcomes.
Basic Science
Detailed examinations of the effects of the flow rate in IT drug
delivery have been conducted by the late C. Bernards in a porcine
model (20). This model has a number of advantages over previous
models in that it mimics IT therapy, namely a closed model with no
CSF loss or disturbance due to sampling, preservation of the effect
of cardiac and respiratory cycles, an upright position and the ability
to study drug concentrations in both spinal fluid and spinal cord.
Bernards compared flow rates of 20–1000 mcL/hour and a bolus
group receiving a bolus of 1000 mcL/hour administered more than
5 min every hour (20). These rates were chosen to be representative
of regular and maximum clinical pump flow rate as well maximum
speeds achieved by bolus administration. The most prominent find-
ing was the limited distribution of bupivacaine and baclofen from
the site of administration, especially in the 20 mcL/hour group. For
both bupivacaine and baclofen, most of the drug recovered in the
CSF and spinal cord in this group was found within 1 cm of the site
of administration. Diffusion of both bupivacaine and baclofen in CSF
or spinal cord parenchyma was increased in the 1000 mcL/hour and
bolus groups compared with the 20 mcL/hour group. Evidence of
greater distribution comes from the dose-normalized CSF area under
the curve and spinal cord concentration data. Evidence that the
bolus group achieved better drug distribution than did the 1000-
mcL/hour group was more subtle but still present. Bernards con-
cluded that CSF is a poorly mixed medium, that CSF motion is
limited and that the spread of drug molecules in CSF is largely
dependent on the kinetic energy imparted to the drug molecules by
121
the infusing mechanism. The clinical implications of Bernards’ work
 DEER ET AL.
Table 27. Recommendations Regarding Infusion Rate by the PACC Using USPSTF Criteria.
Statements
Rate of dispersion in cerebrospinal fluid cannot be attributed to diffusion alone.
Flow rate may not impact analgesia.
Bolus dosing may improve analgesia.
are that the location of the tip of the catheter relative to the tar-
geted spinal cord segment may be critical given the limited capacity
for CSF to distribute drugs away from the catheter tip. Limited distri-
bution of morphine away from the catheter tip may predispose to
formation of IT granulomas, and development of methods to
improve drug distribution may decrease the risk of granuloma
formation.
Flack et al. went on to confirm these findings in a chronic ambula-
tory pig model receiving IT morphine over a period of 14 days (90).
In this experiment, four pigs were implanted with IDD systems and
an infusion of morphine delivered at 20 mcL/hour. The authors con-
cluded that the chronic data confirmation of limited CSF distribution
in the ambulatory animal may help to explain clinical situations in
which a drug delivered as an IT bolus at trial stage is effective at
relieving patient symptoms but, when delivered at the very slow
infusion rates used for chronic infusion, was not effective (i.e., phar-
macokinetic failures of chronic IT drug delivery). Specifically,
although bolus IT injection produces relatively widespread drug dis-
tribution, that was not the case with chronic infusions.
Clinical Studies
Two clinical studies have examined the effect of varying the infu-
sion rate for baclofen and varied mixtures of analgesics, respectively.
Both studies utilized a similar double-blind crossover design over
two-week periods with a constant daily drug dosage throughout
but an infusion rate varying at random from the patient’s baseline
rate to twice and four times this value. In the baclofen study van der
Plas et al. (258) randomized patients who experienced no beneficial
response or excessive side effects to intrathecal baclofen (ITB) infu-
sion for dystonia to either slower infusion rate delivery or four-times
faster infusion rate delivery (FIRD) for two weeks. Patients crossed
over after a one-week washout period. The authors observed no sig-
nificant differences between the two groups for the median change
of numeric rating scale dystonia (20.3 [interquartile range {IQR}
21.1 to 0.5]), pain (0.1 [IQR 20.8 to 1.3]). However, they found that
the frequency of AEs was significantly higher during FIRD (12 vs. 2).
Only patients who were included because side effects to ITB pre-
vented dose escalation preferred FIRD. Investigators concluded that
given a fixed daily dose, a four-times higher infusion rate enhances
the IT distribution of baclofen as evidenced by the significantly
higher number of AEs. However, in CRPS a fourfold higher infusion
rate was not associated with clinically overt improvement of dysto-
nia or pain. Patients in whom side effects restricted further dose
escalations of ITB favored the faster rate because of subjective
improvement of dystonia and pain. Therefore, the utility of a faster
rate of delivery should be further investigated for this group.
To date no clinical studies have compared bolus to continuous
infusion, although the authors understand that such an experiment
is in progress (282).
The use of bolus dosing in addition to continuous flow is possible
with commercially available IDD systems. One prospective registry
of 168 patients suggested that patient-controlled bolus therapy
122
with concomitant constant infusion resulted in improved patient
www.neuromodulationjournal.com C 2017 International Neuromodulation Society
V Neuromodulation 2017; 20: 96–132
Evidence levels Recommendation strength Consensus strength
II-2 B Strong
II-2 C Weak
II-2, II-3 B Strong
satisfaction and reduced need for oral medication supplementation
(241). Recently, it has been reported that patient-controlled IT anal-
gesia with bolus dosing results in better patient satisfaction in
cancer-related pain (282). No further work on bolus dosing with IDD
has been done to our knowledge (283–285).
All told, these data suggest a lack of benefit from increasing infu-
sion rate and, in fact, there may be a clinically significant deleterious
effect, as decreased quality of life has been reported with increasing
flow rate (283). The one preliminary report using intermittent bolus
dosing in addition to constant infusion suggests a positive impact
on patient outcomes but has not been replicated. Reck et al. dem-
onstrated, in a blinded crossover study of ten patients comparing
bolus to continuous infusion, a statistically significant reduction of
numerical rating scores with intermittent, programmed, bolus deliv-
ery, compared to continuous infusion (286). No conclusions regard-
ing safety and efficacy can be drawn from the limited data currently
published. However, with the technology now available making it
possible for bolus therapy in multiple applications to be provided to
patients, this therapy is being utilized by increasing numbers of pro-
viders and data will be forthcoming. Caution and a conservative
approach is advised when choosing to utilize intrathecal bolus ther-
apy as our understanding of flow dynamics, oscillatory mechanisms
and drug bioavailability are still evolving.
The PACC recommendations for infusion rate appear in Table 27.
Baseline Dose of Opioids: High vs. Low or None
The impact of oral opioid therapy on subjects trialed and
implanted for IT therapy has been examined in several recent stud-
ies (14,15,100,198,284,285). The techniques surrounding manage-
ment of oral opioid therapy in those considering IDD range from
leaving the patient on oral opioids and adding IDD to pretrial/
implant taper of opioid medications. Anderson et.al. in 2003
reported outcomes after taking subjects off opioids 12 hours before
the trialing period (284). This was followed by several case studies
that described various methods of tapering mediations during the
trialing period (101,195). Shaladi et al. (198) lowered oral opioid
doses as IT doses increased during the trial, while Kim et al. (100) dis-
continued opioids 4–12 hours prior to trialing.
Subsequent larger studies examined the role of eliminating oral
opioids and the effects on the efficacy of IDD (14,15). Grider et al. in
a small case study and later in a larger retrospective study reported
discontinuing oral opioids for six weeks prior to trial/implant. In that
study the pretrial VAS score on oral opioids was compared to the
VAS following opioid taper and six weeks in an opioid-free state,
demonstrating that patient-recorded VAS scores were virtually iden-
tical after discontinuation of oral opioid therapy (14). Hamza et al.
likewise demonstrated analgesic efficacy in subjects trialed and
maintained on low-dose IT opioids, with most subjects dramatically
reducing or eliminating oral opioid use (15). However, it should be
emphasized that studies on microdosing were not controlled or ran-
domized. Such findings underscore the importance of RCTs.
Several studies have examined the impact of pretrial opioid use
on postimplant IT analgesia. Kim et al. found that pretrial systemic

Table 28. Recommendations Regarding Patient Characteristics Affecting Intrathecal Therapy—Baseline Dose of Opioids, High vs. Low, by the PACC Using
USPSTF Criteria.
Statements
Pretrial opioid dose does not appear to be predictive of
intrathecal drug delivery outcomes.
Effective sustainable analgesia is achievable with intrathecal drug delivery
without systemic supplementation.
opioid requirement was a poor predictor of IT dose, efficacy, or need
to change medication at one year postimplant (100). This study did
find that the trial IT opioid dose was a good predictor of success
with IT therapy. Likewise, Mekhail et al. reported no link between
systemic opioid requirement and efficacy with IT opioids (108). A
recent analysis has suggested a significant cost benefit to the elimi-
nation of oral opioid therapy in those transitioning to IDD (103).
Taken together, these data suggest that the goal of limiting or
eliminating oral opioid therapy in those transitioning to IDD can be
accomplished in many ways. No large-scale trial has compared pre-
implant opioid cessation with postimplant cessation, however, in
the two largest studies to date analgesic efficacy was achieved with
both methods, albeit at lower doses in the pretrial taper study. There
appears to be little value in the preimplantation opioid dose as a
predictor of success with IT therapy (100,108). The impact on
response to ziconotide based on preimplant opioids has not been
determined.
The PACC recommendations regarding patient characteristics that
affect IT therapy appear in Table 28.
PSYCHOLOGICAL CONSIDERATIONS
The general belief that identifying comorbid psychological factors,
which could compromise treatment success, was borrowed from
neurostimulation practice and guidelines and applied to IT therapy,
especially in the noncancer pain setting. Nelson et al. in 1996 (287)
proposed a list of “red flags” to success of treatment that included
suicidality, alcohol or drug dependency, unresolved compensation/
legal issues, severe depression, and so on, which, although not
empirically derived, made sense clinically. This spurred a “rule-out”
approach to the assessment for neuromodulation in general. More
recent guidelines (56) have emphasized the assessment of positive
characteristics such as proper expectation, social support, effective
coping skills, and so on, and the importance of using psychological
intervention before and after internalization of an IDD device. Some
third-party payers mandate this screening process for authorization
of the procedure.
Four questions summarize the practical considerations related to
psychological assessment for IT interventions: 1) Should psychologi-
cal evaluation be performed? 2) If so, when is the best time for eval-
uation? 3) Who should perform the psychological evaluation?
4) What are the best practice guidelines for psychological evalua-
tion? These and other aspects of psychological screening are dis-
cussed more thoroughly in a PACC companion article on screening
trials for IT therapy (118).
A review of the published IT literature from 1998 through 2010
reveals few psychological evaluations in the studies identified (288).
Furthermore, there appear to be few, if any, systematic studies with
sufficient follow-up to determine the contribution of psychological
evaluation to outcome. There has also been criticism of reliance on
www.neuromodulationjournal.com C 2017 International Neuromodulation Society
V Neuromodulation 2017; 20: 96–132
INTRATHECAL THERAPY BEST PRACTICES AND GUIDELINES
Evidence levels Recommendation strength Consensus strength
II-3 C Moderate
II-2 B Moderate
psychological assessment as a component of the selection criteria
(289). The continuing emphasis is on identification of predictive
characteristics. Yet identification of patient states or traits that pre-
dict outcome is not scientifically valid. A more reliable approach is
to assess for and identify psychological symptoms (e.g., depression)
and/or psychiatric diagnoses (e.g., post-traumatic stress disorder),
which could be barriers to a positive outcome.
As noted already, the approach to a patient with cancer-related
pain should be somewhat more flexible. For patients with signifi-
cantly compromised life expectancy (Category 1, Table 7), psycho-
logical evaluation should be considered optional. We encourage
psychological evaluation for patients in whom the disease process
has been arrested but there is a significant probability of reoccur-
rence (Category 2, Table 7). Patients whose cancer has been eradi-
cated by appropriate therapies and continue to manifest chronic
pain secondary to medical treatment/anatomical/disease-related
damage, but wherein life expectancy is only minimally compro-
mised, should be considered in the same context as chronic non-
cancer pain patients (Category 3, Table 7).
Consensus Point 24. Psychological assessment, counseling, and
after care are recommended in appropriate candidates. The use of
an assessment is critical in all noncancer patients receiving an intra-
thecal drug delivery system. An extensive discussion of the proper
tools and techniques of this screening process is presented in the
PACC trialing recommendations (118).
Consensus Point 25. Psychological screening is not required for
end-of-life patients, but psychological counseling should be
considered.
EDUCATIONAL REQUIREMENTS FOR
IMPLANTING AND/OR MANAGING IDDS
THERAPY
The extensive scope and breadth of this sixth edition of the PACC
guidelines is a reflection of the growth of knowledge related to the
safe implementation of implantable IT therapies. In addition to the
rapid growth of the preclinical and clinical science knowledge, there
has also been an increase in the number of commercially available
implantable IDDSs. While all of these devices function by pumping
medication from an implantable reservoir to the IT space via an
implanted catheter, their propellant mechanism, MRI compatibility,
and device-specific engineering limitations vary substantially. Rapid-
ly advancing clinical and scientific knowledge combined with the
variability of pump designs and function make it imperative that
providers throughout the health care continuum are thoroughly
trained and credentialed to provide appropriate and safe care to
patients implanted with these devices. The specific training recom-
mendations are largely dependent on the scope of practice of the
123
individual provider, the disease state being treated, and the device
 DEER ET AL.
being used. Every patient should have a comprehensive pump man-
ager who is accessible throughout the continuum of medical care.
The NANS in partnership with the INS is focused on education and
credentialing as a strategy to continue to improve safety and effica-
cy of these therapies by promoting improved implementation and
maintenance training and assessment. Credentialing by the hospital
site of service is currently the mechanism of certification, and inher-
ently this is nonuniform.
First Responders
All health care providers, and especially first responders to disease
or trauma, as part of their core curriculum and training, should be
trained to identify patients with implanted systems. They should be
aware that there are multiple different types of implanted systems,
including pacemakers, SCS devices and implanted pumps, and
should be able to distinguish between them. When an implanted
infusion pump is suspected (most often large circular device vs. neu-
rostimulator, pacemaker, or other smaller implanted device), the first
responder should have information available to them and be trained
to contact a pump manager or a representative of the pump
manufacturer.
Suggested Training Milestones
Suggested training regarding implanted neuromodulation devi-
ces, their indications and the processes for monitoring and refilling
them, should be performed in nursing schools, medical schools and
schools that train allied health professionals. Likewise, this same
information should be provided in continuing medical education
courses for further education of these health care providers. Current-
ly most programs in Europe, Asia, Australia, and the United States
do not include this information.
Consensus Point 26. The PACC recommends an overview of
intrathecal drug delivery be added to the basic curriculum of physi-
cians in training, nurses, and allied health care professionals.
Pump Interrogator
All patients admitted to a medical facility should have their IDDS,
medications, and daily dosage of IT medication documented as part
of the medical record. For programmable pumps this often requires
an electronic interrogation of the pump and documentation of the
drug concentrations and delivery modes. Documentation of type
and manufacturer of pump identification for drug delivery should
be considered the standard of care prior to performing an MRI or
pursuing elective surgery, and is imperative for all patients admitted
to a hospital. The interrogator of the IDDS is expected to be appro-
priately trained to perform the following:
•
Data gathering and communication skill set
• granuloma formation, and signs and symptoms of medication over/
Interrogate an IDDS without altering programming
• Determine manufacturer, model, and when the device was
implanted
• Determine the drug(s) and dosage(s) delivered in a day
• Determine the drug refill alarm dates
• Access and interpret alarm logs
• Contact a physician pump manager
Suggested Training Milestones and Credentialing for Pump
Interrogators
Detailed didactic lectures reviewing all currently available devices
and techniques for pump interrogation are essential. These lectures
124
should include the x-ray imaging of all commercially available
www.neuromodulationjournal.com C 2017 International Neuromodulation Society
V Neuromodulation 2017; 20: 96–132
devices to better identify them for patients who are unable to pro-
vide detailed information regarding their pump. For credentialing
purposes, a minimum of ten interrogations per year for each make is
recommended prior to independently interrogating a pump without
the assistance of an industry representative or another supervising
provider. These ten interrogations can be performed in one setting
as part of a hands-on training workshop. The PACC feels that most
programs do not currently meet these standards.
Consensus Point 27. The PACC recommends each accredited
facility have the ability to evaluate an indwelling implanted device,
including pump and catheter system. We encourage manufacturers
and facilities to collaborate on this important issue with a goal of
meeting compliance by the next scheduled PACC in 2019.
Personnel Who Perform Maintenance and Programming/
Reservoir Refill of IDDS
Virtually all patients implanted with an IDDS will need pump inter-
rogation and refill at regular intervals, whether they are inpatients,
outpatients, or homebound patients. The provider of this service
may or may not be the managing physician but should be under
the supervision of a managing physician. The provider of this service
is expected to be appropriately trained to perform the following
and have a supervising IDDS physician pump manager. In many set-
tings these tasks are accomplished by licensed nursing infusion
company employees. These service providers should be properly
trained and supervised.
Programming and Refilling Skill Set
• Perform pump interrogation and programming
• Safely perform an aseptic pump refill (preferably with and without
image guidance)
• Diagnose and detect a pocket fill and notify a credentialed physi-
cian pump manager for management
• Identify residual volume discrepancies and notify a credentialed
physician pump manager for management
• Be familiar with medication formulations that are appropriate
and inappropriate for intrathecal drug delivery (PACC
recommendations)
• Have supervision by a credentialed physician pump manager
Suggested Training Milestones
Training should include the detailed didactic lectures described
above and should also include added lectures on the medication
choices for IT therapy and pharmacodynamics and pharmacokinetics
of IT drug delivery. Additional didactic lectures on the indications
and contraindications for implementing IDDS therapy as well as
evaluating and recognizing serious AEs such as pump pocket fills,
underdosage should apply. For training purposes, a minimum of 20
supervised pump refills is recommended for initial assessment prior
to refills being performed independently. A minimum of ten pump
refills per year is suggested to by consensus in order that physician
pump managers maintain clinical competency for each make of
device. If these minimums are not achievable for every make, the
pump refill should be directly supervised by a physician pump man-
ager, a credentialed nurse who has met these requirements, or by a
manufacturer representative.
Consensus Point 28. The training of all personnel for device eval-
uation and refilling is an important part of patient care. This training
should be device specific and supervised carefully. The PACC

recommends that 20 refills be supervised before independent prac-
tice is approved. Two or more trained individuals should check all
reprogramming.
IDDS Implanting Physician
Intrathecal pump implantation and/or explantation requires basic
surgical skills necessary to perform the procedure safely. Not all
implanters are IDDS managers and not all mangers are IDDS implan-
ters. In those cases, where the manager and implanter are not the
same individual, the two providers should be in close communica-
tion regarding appropriate planning for the placement of the pump
and catheter and to determine the concentrations and types of
medications to be infused by the pump. In addition, prior to implan-
tation, a designated manager must be identified and available to
manage the pump immediately after implantation and to monitor
for adverse side effects of IT medications. In cases of explantation,
the manager should inform the explanter of the systemic medica-
tion to be delivered once the IDDS is removed. The IDDS implanter
is expected to be appropriately trained to perform the following
procedures and have a supervising IDDS physician pump manager
available before the case to address potential drug overdose and
underdose.
Implanter Skill Set
• Appropriately place a pump subcutaneously and tunnel a catheter
from the pump site to the catheter insertion site
• Appropriate and complete medical training in the area of Pain
Medicine or Surgery, with a focus on implantable technologies,
recognized by the country and area of practice.1
• Diagnose and troubleshoot potential intraoperative surgical
complications
• Diagnose and troubleshoot potential postoperative and chronic
complications of IDDS therapy requiring surgical intervention, for
example, granuloma formation, catheter breaks, catheter
occlusions
Suggested Training Milestones
The implanter should have also completed the basic didactic
training required of the refill provider as previously described. A
minimum of ten supervised implant and/or explant cases should be
performed under supervision of a credentialed implanter. Another
five cases over the course of two years should be completed to
maintain clinical competency. For intracerebroventricular placed
catheters, implanter must be credentialed in neurosurgery.
1
Special comment is necessary regarding the suggested formal medical
and surgical education. As this manuscript serves as a living, international
document, it is clear that no uniform credentialing body exists to measure
(or test) specific training criteria over such a diverse group. However, basic
skill standards can be measured. Each implanter must undergo appropri-
ate surgical tissue management training, with specific experience with
implanting IT therapy. Internationally, the World Institute of Pain (WIP) cre-
ated an exam to standardize internationally delivered interventional pain
management, and there is discussion surrounding this effort through the
educational committee collaborations of NANS, INS, and WIP. In the United
States, since the inception of an American Council of Graduate Medical Educa-
tion (ACGME) certified training program in Pain Medicine and surgical subspe-
cialties of Neurosurgery or Orthopedic Spine Surgery, it is recommended that
implanters have underdone and completed such training. This recommenda-
tion, does not, however, impact “legacy or grandfathered” practitioners for
whom no such training was available.
www.neuromodulationjournal.com C 2017 International Neuromodulation Society
V Neuromodulation 2017; 20: 96–132
INTRATHECAL THERAPY BEST PRACTICES AND GUIDELINES
Consensus Point 29. In order to offer intrathecal therapies
regardless of primary specialty, the physician should be supervised
in a minimum of ten implant and/or explant cases.
Consensus Point 30. In order to maintain skills, the implanter
should be involved in five cases more than two years, or should
undergo additional hands-on certified educational training to refresh
skills.
Physician Pump Manager
All patients with IT pumps should have a designated physician
pump manager. This individual will provide the coordination of care
amongst other providers across the health care spectrum.
Pump Manager Skill Set
• Be a leader who can manage therapy and complications associat-
ed with IDDS
• Diagnose and manage overdose and underdose (withdrawal
syndrome)
• Perform dye studies, rotor studies, interpret imaging (fluoroscopy,
CT, MRI)
• Understand the pharmacology of IT drug delivery
• Responsibilities include, but are not limited to: patient and device
selection, pump fills, prescribing, altering therapy, interactions
with non-IT medications and systems (e.g., MRI)
Suggested Training Milestones
If the manager is also the implanter this individual should meet all
the requirements previously described, in addition to being compe-
tent at performing dye studies/side access port procedures for each
make of pumps prior to performing management services indepen-
dently. For specific manufactured devices where fewer than ten
cases have been completed, another manager with skills in that
device should proctor or assistance be obtained from the manufac-
turer. In addition to the information provided in the didactic lectures
previously listed, the manager should also be knowledgeable about
all of the advanced intrathecal pharmacokinetic and pharmacody-
namic principles and current medical guidelines. This same knowl-
edge base is required for the managers who are not implanters.
Consensus Point 31. The training to manage an intrathecal
device is critical to the long-term success of the therapy. All pump
management physicians should have ongoing educational training
that includes knowledge of all current and FDA-approved devices
and future devices approved by regulatory bodies for research, and
found to be clinically relevant. Managing physicians who are not
implanting physicians are expected to be trained at the same level
as those who both implant and manage devices.
Consensus Point 32. To establish a national data base for all
intrathecal pump-managing physicians or healthcare professionals
as a repository of current pump settings, mediations, efficacy, and
side effects.
CONCLUSIONS
The previous PACC work led to improved patient safety and effi-
cacy and advanced questions that fostered additional IT drug
research. In the same spirit, this present manuscript presents the
next step in algorithmic thinking. The creation of new algorithmic
tracks for neuropathic and nociceptive pain is an important step in
125
improving patient care. The panel encourages continued research
 DEER ET AL.
and development, including the development of new drugs, devi-
ces, and safety recommendations to improve the care of patients
whom we strive to help. The PACC is hopeful that the time interval
between now and our next update produces new insights in the
field of IT drug delivery.
The creation of this consensus statement has depended heavily
on available literature, clinical experience and scientific discourse.
Despite our mechanisms used to create the best consensus recom-
mendations possible, the final conclusions include a subjective com-
ponent and may be controversial. The panel has addressed
nociceptive and neuropathic pain pathways to best treat pain by IT
infusion. The panel had considered a third pathway for mixed pain
syndromes but considering the heterogeneous components of this
complex patient group, the reader is advised to use best clinical
judgment to choose the most appropriate pathway, with the realiza-
tion that the patient may exhibit different components of pain at dif-
fering and various times.
Acknowledgement
The PACC was initiated by INS and funded by unrestrict-
ed educational grants from Medtronic Inc., and Jazz Phar-
maceuticals, Inc. No corporate entities had any direct input
into the contents of the manuscript or the conclusions of
the collaborators. Sarah Staples, MA, ELS, assisted with
manuscript preparation.
Authorship Statements
Dr. Deer served as primary author, project organizer and editor;
Dr. Doleys, Falowski, Jacobs, Kim, and Narouze, performed literature
searches; Drs. Deer, Hayek, Pope, Grider, and Erdek prepared evi-
dence tables; Drs. Huntoon, Mekhail, and Krames served as senior
editors; the remaining authors acquired or interpreted data, wrote
sections of the manuscript, and provided critical reviews and
editing.
How to Cite this Article:
Deer T.R., Pope J.E., Hayek S.M., Bux A., Buchser E., Eldabe
S., De Andres J.A., Erdek M., Patin D., Grider J.S., Doleys
D.M., Jacobs M.S., Yaksh T.L., Poree L., Wallace M.S.,
Prager J., Rauck R., DeLeon O., Diwan S., Falowski S.M.,
Gazelka H.M., Kim P., Leong M., Levy R.M., McDowell II G.,
McRoberts P., Naidu R., Narouze S., Perruchoud C., Rosen
S.M., Rosenberg W.S., Saulino M., Staats P., Stearns L.J.,
Willis D., Krames E., Huntoon M., Mekhail N. 2017. The
Polyanalgesic Consensus Conference (PACC):
Recommendations on Intrathecal Drug Infusion Systems
Best Practices and Guidelines.
Neuromodulation 2017; 20: 96–132
REFERENCES
1. Pope JE, Deer TR. Intrathecal pharmacology update: novel dosing strategy for
intrathecal monotherapy ziconotide on efficacy and sustainability. Neuromodula-
tion 2015;18:414–420.
2. Bennett G, Burchiel K, Buchser E et al. Clinical guidelines for intraspinal infusion:
126
report of an expert panel. PolyAnalgesic Consensus Conference 2000. J Pain Symp-
tom Manage 2000;20:S37–S43.
www.neuromodulationjournal.com C 2017 International Neuromodulation Society
V Neuromodulation 2017; 20: 96–132
3. Pope JE, Falowski S, Deer TR. Advanced waveforms and frequency with spinal cord
stimulation: burst and high frequency energy delivery. Expert Rev Med Devices
2015;12:431–437.
4. Consent Decree of Permanent Injunction. United States of America vs. Medtronic,
Inc., and S. Omar Ishrak and Thomas M. Teft. United States District Court for the
District of Minnesota. April 2015.
5. Pope J, Poree L, McRoberts WP, Falowski S, Deer T. Consent decree: physician and
institution ramifications? Neuromodulation 2015;18:653–656.
6. Hassenbusch SJ, Portenoy RK, Cousins M et al. Polyanalgesic Consensus Confer-
ence 2003: an update on the management of pain by intraspinal drug delivery—
report of an expert panel. J Pain Symptom Manage 2004;27:540–563.
7. Deer T, Krames ES, Hassenbusch SJ et al. Polyanalgesic Consensus Conference
2007: recommendations for the management of pain by intrathecal (intraspinal)
drug delivery: report of an interdisciplinary expert panel. Neuromodulation 2007;
10:300–328.
8. Deer TR, Prager J, Levy R et al. Polyanalgesic Consensus Conference 2012: recom-
mendations for the management of pain by intrathecal (intraspinal) drug delivery:
report of an interdisciplinary expert panel. Neuromodulation 2012;15:436–464; dis-
cussion 464–436.
9. Hayek SM, Hanes MC. Intrathecal therapy for chronic pain: current trends and
future needs. Curr Pain Headache Rep 2014;18:388.
10. Deer TR, Pope JE. Factors to consider in the choice of intrathecal drug in the treat-
ment of neuropathic pain. Expert Rev Clin Pharmacol 2015;8:507–510.
11. Smith TJ, Staats PS, Deer T et al. Randomized clinical trial of an implantable drug
delivery system compared with comprehensive medical management for refracto-
ry cancer pain: impact on pain, drug-related toxicity, and survival. J Clin Oncol
2002;20:4040–4049.
12. Hayek SM, Veizi IE, Narouze SN, Mekhail N. Age-dependent intrathecal opioid esca-
lation in chronic noncancer pain patients. Pain Med 2011;12:1179–1189.
13. Dominguez E, Sahinler B, Bassam D et al. Predictive value of intrathecal narcotic tri-
als for long-term therapy with implantable drug administration systems in chronic
non-cancer pain patients. Pain Pract 2002;2:315–325.
14. Grider JS, Harned ME, Etscheidt MA. Patient selection and outcomes using a low-
dose intrathecal opioid trialing method for chronic nonmalignant pain. Pain Physi-
cian 2011;14:343–351. 3
15. Hamza M, Doleys D, Wells M et al. Prospective study of 3-year follow-up of low-
dose intrathecal opioids in the management of chronic nonmalignant pain. Pain
Med 2012;13:1304–1313.
16. Bernards CM. Recent insights into the pharmacokinetics of spinal opioids and the
relevance to opioid selection. Curr Opin Anaesthesiol 2004;17:441–447.
17. Krames ES. Intrathecal infusional therapies for intractable pain: patient manage-
ment guidelines. J Pain Symptom Manage 1993;8:36–46.
18. Friese S, Hamhaber U, Erb M, Kueker W, Klose U. The influence of pulse and respi-
ration on spinal cerebrospinal fluid pulsation. Investigat Radiol 2004;39:120–130.
19. Henry-Feugeas MC, Idy-Peretti I, Baledent O et al. Origin of subarachnoid cerebro-
spinal fluid pulsations: a phase-contrast MR analysis. Magn Reson Imaging 2000;18:
387–395.
20. Bernards CM. Cerebrospinal fluid and spinal cord distribution of baclofen and
bupivacaine during slow intrathecal infusion in pigs. Anesthesiology 2006;105:169–
178.
21. Doleys DM. Psychologic evaluation for patients undergoing neuroaugmentative
procedures. Neurosurg Clin North Am 2003;14:409–417.
22. Doleys DM, Brown J. MMPI profile as an outcome “predictor” in the treatment of
noncancer pain patients utilizing intraspinal opioid therapy. Neuromodulation
2001;4:93–97.
23. Doleys DM, Brown JL, Ness T. Multidimensional outcomes analysis of intrathecal,
oral opioid, and behavioral-functional restoration therapy for failed back surgery
syndrome: a retrospective study with 4 years’ follow-up. Neuromodulation 2006;9:
270–283.
24. Product Surveillance Registry (PSR) Database. Minneapolis, MN: Medtronic plc;
2015.
25. Harris RP, Helfand M, Woolf SH et al. for the Methods Work Group, Third US Pre-
ventive Service Task Force. A review of the process. Am J Prev Med 2001;20:21–35.
26. Hayek SM, Deer TR, Pope JE, Panchal S, Patel V, Burton AW. Intrathecal therapy for
cancer and non-cancer pain. Pain Physician 2011;14:219–248.
27. Prager J, Deer T, Levy R et al. Best practices for intrathecal drug delivery for pain.
Neuromodulation 2014;17:354–372; discussion 372.
28. Deer TR, Smith HS, Cousins M et al. Consensus guidelines for the selection and
implantation of patients with noncancer pain for intrathecal drug delivery. Pain
Physician 2010;13:E175–E213.
29. Falco F, Patel V, Hayek S et al. Intrathecal infusion systems for long-term manage-
ment of chronic cancer pain: an update of assessment of evidence. Pain Physician
2013;16:SE185–SE216.
30. Kloth D. President’s Message. NANS Newsletter 2015;11:1–2.
31. Yaksh TL. New horizons in our understanding of the spinal physiology and phar-
macology of pain processing. Semin Oncol 1993;20(2Suppl.1):6–18.
32. Rauck RL, Wallace MS, Leong MS et al. A randomized, double-blind, placebo con-
trolled study of intrathecal ziconotide in adults with severe chronic pain. J Pain
Symptom Manage 2006;31:393–406.
33. Staats PS, Yearwood T, Charapata SG et al. Intrathecal ziconotide in the treatment
of refractory pain in patients with cancer or AIDS: a randomized controlled trial.
JAMA 2004;291:63–70.
34. Wallace MS, Charapata SG, Fisher R et al. Intrathecal ziconotide in the treatment of
chronic nonmalignant pain: a randomized, double-blind, placebo-controlled clini-
cal trial. Neuromodulation 2006;9:75–86.

35. Pope JE, Deer TR, McRoberts WP. Intrathecal therapy: the burden of being posi-
tioned as a salvage therapy. Pain Med 2015;16:2036–2038.
36. Poree L, Krames E, Pope J, Deer TR, Levy R, Schultz L. Spinal cord stimulation as
treatment for complex regional pain syndrome should be considered earlier than
last resort therapy. Neuromodulation 2013;16:125–141.
37. Deer TR, Caraway D, Wallace M. A definition of refractory pain to help determine
suitability for device implantation. Neuromodulation 2014;17:711–715.
38. Pope JE, Deer TR. Intrathecal drug delivery for pain: a clinical guide and future
directions. Pain Mange 2015;5:175–183.
39. Centers for Disease Control (CDC). Vital signs: overdoses of prescription opioid
pain relievers—United States, 1999-2008. Morb Mortal Wkly Rep 2011;60:1–6.
40. Han B, Compton WM, Jones JM et al. Nonmedical prescriptive opioid use and use
disorders among adults aged 18-64 years in the United States 2003-2013. JAMA
2015;314:1468–1478.
41. Coffey RJ, Owens ML, Broste SK et al. Mortality associated with implantation and
management of intrathecal opioid drug infusion systems to treat noncancer pain.
Anesthesiology 2009;111:881–891.
42. Coffey RJ, Owens ML, Broste SK et al. Medical practice perspective: identification
and mitigation of risk factors for mortality associated with intrathecal opioids for
non-cancer pain. Pain Med 2010;11:1001–1009.
43. Prager J, Jacobs M. Evaluation of patients for implantable pain modalities: medical
and behavioral assessment. Clin J Pain 2001;17:206–214.
44. North RB, Kidd DH, Farrokhi F, Piantadosi SA. Spinal cord stimulation versus repeat-
ed lumbosacral spine surgery for chronic pain: a randomized, controlled trial. Neu-
rosurgery 2005;56:98–106; discussion 106–107. Available from: http://www.ncbi.
nlm.nih.gov/pubmed/15617591.
45. Kumar K, Taylor RS, Jacques L et al. Spinal cord stimulation versus conventional
medical management for neuropathic pain: a multicentre randomised controlled
trial in patients with failed back surgery syndrome. Pain 2007;132:179–188.
46. Kumar K, Hunter G, Demeria D. Spinal cord stimulation in treatment of chronic
benign pain: challenges in treatment planning and present status, a 22-year expe-
rience. Neurosurgery 2006;58:481–496.
47. Hayek SM, Veizi HM. Treatment-limiting complications of percutaneous spinal cord
stimulator implants: a review of eight years of experience from an academic center
database. Neuromodulation 2015;18:603–609.
48. Al-Kaisy A, Van Buyten J-P, Smet I et al. Sustained effectiveness of 10 kHz high-
frequency spinal cord stimulation for patients with chronic, low back pain: 24-
month results of a prospective multicenter study. Pain Med 2014;15:347–354.
49. deVos CC, Dijkstra C, Lenders MW, Holsheimer J. Spinal cord stimulation with
hybrid lead relieves pain in low back and legs. Neuromodulation 2012;15:118–123.
50. Liem L, Russo M, Huygen FJ et al. One-year outcomes of spinal cord stimulation of
the dorsal root ganglion in the treatment of chronic neuropathic pain. Neuromodu-
lation 2015;18:48–49.
51. Rosenberg J, Tavel E, Jackson A et al. Development of new pain affects the long-
term patient experience with spinal cord stimulation. Presented at the 7th World
Congress World Institute of Pain; May 2014; Maastricht, The Netherlands. (WIP-
0412).
52. Saulino M, Stearns L, (moderators). Intrathecal Therapies: Basic Challenges. Concur-
rent session presented at: North American Neuromodulation Society 19th Annual
Meeting; Dec. 2015; Las Vegas, NV.
53. Lindblom U, Meyerson BA. Influence on touch, vibration and cutaneous pain of
dorsal column stimulation in man. Pain 1975;1:257–270.
54. Anderson F, Downing GM, Hill J, Casorso L, Lerch N. Palliative performance scale
(PPS): a new tool. J Palliative. Care 1996;12:5–11.
55. ECOG-ACRIN. ECOG performance status. http://ecog-acrin.org/resources/ecog-per-
formance-status.
56. Deer TR, Smith T, Burton AW. Comprehensive consensus based guidelines on
intrathecal drug delivery systems in the treatment of pain caused by cancer pain.
Pain Physician 2011;14:E283–E312.
57. Codman, a division of Johnson & Johnson. Implantable infusion pumps. http://
www.codmanpumps.com/Products_pumps_overview.asp.
58. Medtronic plc. Targeted drug delivery. Indications, safety and warnings. Synchromed
II. http://professional.medtronic.com/pt/neuro/idd/ind/index.htm#.VorVwVKkx-A.
59. Deer TR, Prager J, Levy R et al. Polyanalgesic Consensus Conference–2012: recom-
mendations on trialing for intrathecal (intraspinal) drug delivery: report of an inter-
disciplinary expert panel. Neuromodulation 2012;15:420–435; discussion 435.
60. Deer TR, Prager J, Levy R, et al. Polyanalgesic Consensus Conference–2012: consen-
sus on diagnosis, detection, and treatment of catheter-tip granulomas (inflamma-
tory masses). Neuromodulation 2012;15:483–495; discussion 496.
61. Deer TR, Levy R, Prager J et al. Polyanalgesic Consensus Conference–2012: recom-
mendations to reduce morbidity and mortality in intrathecal drug delivery in the
treatment of chronic pain. Neuromodulation 2012;15:467–482; discussion 482.
62. Ver Donck A, Vranken JH, Puylaert M et al. Intrathecal drug administration in
chronic pain syndromes. Pain Pract 2014;14:461–476.
63. Webster LR, Choi Y, Desai H, Webster L, Grant BJ. Sleep-disordered breathing and
chronic opioid therapy. Pain Med 2008;9:425–432.
64. Provenzano DA, Deer T, Luginbuhl Phelps A et al. An international survey to under-
stand infection control practices for spinal cord stimulation. Neuromodulation
2016;19:71–84.
65. Deer TR, Mekhail N, Provenzano D et al. The appropriate use of neurostimulation:
avoidance and treatment of complications of neurostimulation therapies for the
treatment of chronic pain. Neuromodulation 2014;17:571–597.
66. Bowater RJ, Stirling SA, Lilford RJ. Is antibiotic prophylaxis in surgery a generally
effective intervention? Testinga generic hypothesis over a set of meta-analyses.
Ann Surg 2009;249:551–556.
www.neuromodulationjournal.com C 2017 International Neuromodulation Society
V Neuromodulation 2017; 20: 96–132
INTRATHECAL THERAPY BEST PRACTICES AND GUIDELINES
67. Alexander JW, Solomkin JS, Edwards MJ. Updated recommendations for control of
surgical site infections. Ann Surg 2011;253:1082–1093.
68. Forse RA, Karam B, MacLean LD, Christou NV. Antibiotic prophylaxis for surgery in
morbidly obese patients. Surgery 1989;106:750–756; discussion 756–757.
69. Hetem DJ, Bootsma MC, Bonten MJ. Prevention of surgical site infections:
decontamination with mupirocin based on preoperative screening for Staphy-
lococcus aureus carriers or universal decontamination? Clin Infect Dis 2016;62:
631–636.
70. Kohler P, Sommerstein R, Sch€ onrath F et al. Effect of perioperative mupirocin and
antiseptic body wash on infection rate and causative pathogens in patients under-
going cardiac surgery. Am J Infect Control 2015;43:e33–e38.
71. Lamplot JD, Luther G, Mawdsley EL, Luu HH, Manning D. Modified protocol
decreases surgical site infections after total knee arthroplasty. J Knee Surg 2015;28:
395–403. Erratum in: J Knee Surg 2015 Oct;28(5):e1.
72. Anderson MJ, David ML, Scholz M et al. Efficacy of skin and nasal povidone-iodine
preparation against mupirocin-resistant methicillin-resistant Staphylococcus aureus
and S. aureus within the anterior nares. Antimicrob Agents Chemother 2015;59:
2765–2773.
73. Bryce E, Wong T, Forrester L et al. Nasal photodisinfection and chlorhexidine wipes
decrease surgical site infections: a historical control study and propensity analysis.
J Hosp Infect 2014;88:89–95. doi: 10.1016/j.jhin.2014.06.017. Epub 2014 Aug 1. Erra-
tum in: J Hosp Infect 2015;91(1):93.
74. Phillips M, Rosenberg A, Shopsin B et al. Preventing surgical site infections: a ran-
domized, open-label trial of nasal mupirocin ointment and nasal povidone-iodine
solution. Infect Control Hosp Epidemiol 2014;35:826–832.
75. Tanner J, Parkinson H. Double gloving to reduce surgical cross-infection. Cochrane
Database Syst Rev 2006;19:CD003087.
76. Excellence NICE. Surgical site infection: prevention and treatment of surgical site infec-
tion. Clinical Guideline 74. Manchester, UK: National Institute for Health and Clinical
Excellence, 2008.
77. Darouiche RO, Wall MJ, Itani KM et al. Chlorhexidine-alcohol versus povidone-
iodine for surgical-site antisepsis. N Engl J Med 2010;362:18–26.
78. Hutchison JJ, McGuckin M. Occlusive dressings: a microbiologic and clinical review.
Am J Infect Control 1990;18:257–268.
79. Horlocker TT, Wedel DJ, Rowlingson JC et al. Regional anesthesia in the patient
receiving antithrombotic or thrombolytic therapy: American Society of Regional
Anesthesia and Pain Medicine Evidence-Based Guidelines (Third Edition). Reg
Anesth Pain Med 2010;35:64–101.
80. Atchison SR, Durant PA, Yaksh TL. Cardiorespiratory effects and kinetics of intrathe-
cally injected D-Ala2-D-Leu5-enkephalin and morphine in unanesthetized dogs.
Anesthesiology 1986;65:609–616.
81. Ummenhofer WC, Arends RH, Shen DD, Bernards CM. Comparative spinal distribu-
tion and clearance kinetics of intrathecally administered morphine, fentanyl, alfen-
tanil, and sufentanil. Anesthesiology 2000;92:739–753.
82. Ogden CL, Carrol MD, McDowell MA, Flegal KM. Obesity among adults in the Unit-
ed States—no statistically significant change since 2003-2004. NCHS Data Brief
2007;1:1–8.
83. Farney RJ, Walker JM, Cloward TV, Rhondeau S. Sleep-disordered breathing associ-
ated with long-term opioid therapy. Chest 2003;123:632–639.
84. Horlocker TT, Burton AW, Connis RT et al. Practice guidelines for the prevention,
detection, and management of respiratory depression associated with neuraxial
opioid administration. American Society of Anesthesiologists Task Force on Neu-
raxial Opioids. Anesthesiology 2009;110:218–230.
85. Ornstein E, Mateo RS, Effects of opioids. In: McLeskey CH, ed. Geriatric anesthesiolo-
gy, 1st ed. Philadelphia: Williams & Wilkins, 1997:259.
86. Rauck R, Webster L, Wallace M et al. Effect of concomitant antidepressant and anti-
convulsant use on adverse events in patients receiving intrathecal ziconotide in a
long-term extension study. Poster resented at: West Virginia Society of Interven-
tional Pain Physicians; Puerto Rico, 2014.
87. Dworzak H, Fuss F, B€ uttner T. Persisting respiratory depression following intrathe-
cal administration of morphine and simultaneous sedation with midazolam. Anaes-
thesist 1999;48:639–641. German.
88. Rawal N, Tandon B. Epidural and intrathecal morphine in intensive care units.
Intensive Care Med 1985;11:129–133.
89. Kotlinska-Lemieszek A, Klepstad P, Haugen DF. Clinically significant drug-drug
interactions involving opioid analgesics used for pain treatment in patients with
cancer: a systematic review. Drug Des Devel Ther 2015;169:5255–5267.s
90. Flack SH, Anderson CM, Bernards C. Morphine distribution in the spinal cord after
chronic infusion in pigs. Anesth Analg 2011;112:460–464.
91. Flack SH, Bernards CM. Cerebrospinal fluid and spinal cord distribution of hyper-
baric bupivacaine and baclofen during slow intrathecal infusion in pigs. Anesthesi-
ology 2010;112:165–173.
92. Yaksh TL, de Kater A, Dean R, Best BM, Miljanich GP. Pharmacokinetic analysis of
ziconotide (SNX-111), an intrathecal N-type calcium channel blocking analgesic,
delivered by bolus and infusion in the dog. Neuromodulation 2012;15:508–519; dis-
cussion 519.
93. Kabulski J, Poree L, Wolbers L, Quantitative determination & pharmacokinetic
modeling of intrathecally administered opioids in plasma. Poster presented at:
International Neuromodulation Society; June 2015: Berlin, Germany.
94. Fitzgibbon DR, Rathmell JP, Michna E, Stephens LS, Posner KL, Domino KB. Mal-
practice claims associated with medication management for chronic pain. Anesthe-
siology 2010;112:948–956.
95. Dunn KM, Saunders KW, Rutter CM. Overdose and prescribed opioids:
127
associations among chronic noncancer pain patients. Ann Intern Med 2010;152:
85–92.
 DEER ET AL.
96. Chou R, Turner JA, Devine EB. The effectiveness and risks of long-term opioid ther-
apy for chronic pain: a systemic review for a National Institutes of Health Pathways
to Prevention Workshop. Ann Intern Med 2015;162:276–286.
97. Partnership Healthplan of California. http://www.partnershiphp.org/Pages/PHC.aspx.
98. Tyler CB, Advokat C. Investigation of “cross-tolerance” between systemic and intra-
thecal morphine in rats. Physiol Behav 1986;37:27–32.
99. Veizi IE, Hayek SM, Narouze S, Pope JE, Mekhail N. Combination of intrathecal
opioids with bupivacaine attenuates opioid dose escalation in chronic noncancer
pain patients. Pain Med 2011;12:1481–1489.
100. Kim D, Sidov A, Mandhare V, Shuster A. Role of pretrial systemic opioid require-
ments intrathecal trial dose and non-psychological factors as predictors of out-
come of intrathecal pump therapy: one clinician’s experience with lumbar
postlaminectomy pain. Neuromodulation 2011;14:165–175.
101. Grider JS, Etscheidt MA, Harned ME et al. Trialing and maintenance dosing using a
low-dose intrathecal opioid method for chronic nonmalignant pain: a prospective
36-month study. Neuromodulation 2015;19:206–219.
102. Centers for Disease Control. Guideline for Prescribing Opioids for Chronic Pain. http://
www.cdc.gov/drugoverdose/pdf/guidelines_factsheet-a.pdf.
103. Hatheway JA, Caraway D, David G et al. Systemic opioid elimination after implanta-
tion of an intrathecal drug delivery system significantly reduced health-care
expenditures. Neuromodulation 2015;18:207–213; discussion 213.
104. Yaksh TL, Onofrio BM. Retrospective consideration of the doses of morphine given
intrathecally by chronic infusion in 163 patients by 19 physicians. Pain 1987;31:
211–223.
105. Onofrio BM, Yaksh TL. Long-term pain relief produced by intrathecal morphine
infusion in 53 patients. J Neurosurg 1990;72:200–209.
106. Gestin Y, Vainio A, Pegurier AM. Long-term intrathecal infusion of morphine in
the home care of patients with advanced cancer. Acta Anaesthesiol Scand
1997;41:12–17.
107. Roberts LJ, Finch PM, Goucke CR, Price LM. Outcome of intrathecal opioids in
chronic non-cancer pain. Eur J Pain 2001;5:353–361.
108. Mekhail N, Mahboobi R, Farajzadeh Deroee A et al. Factors that might impact intra-
thecal drug delivery (IDD) dose escalation: a longitudinal study. Pain Pract 2014;14:
301–308.
109. Maves TJ, Gebhart GF. Antinociceptive synergy between intrathecal morphine and
lidocaine during visceral and somatic nociception in the rat. Anesthesiology 1992;
76:91–99.
110. Penning JP, Yaksh TL. Interaction of intrathecal morphine with bupivacaine and
lidocaine in the rat. Anesthesiology 1992;77:1186–2000.
111. Prialt (ziconotide) Intrathecal Infusion [package insert]. http://www.prialt.com/.
112. Yaksh TL, Fisher C, Hochman T, Wiese A. Current and future issues in the develop-
ment of spinal agents for the management of pain. Curr Neuropharmacol 2016; in
press.
113. Wallace MS, Rauck RL, Deer T. Ziconotide combination intrathecal therapy: ratio-
nale and evidence. Clin J Pain 2010;26:635–644.
114. de la Calle Gil AB, Pe~na Vergara I, Cormane Bornacelly MA, Pajuelo Gallego A. Intra-
thecal ziconotide and morphine for pain relief: a case series of eight patients with
refractory cancer pain, including five cases of neuropathic pain. Neurol Ther 2015;4:
159–168.
115. Shields DE, Aclan J, Szatkowski A. Chemical stability of admixtures combining zico-
notide with fentanyl or sufentanil during simulated intrathecal administration. Int J
Pharm Compounding 2008;12:463–466.
116. Mohammed SI, Eldabe S, Simpson KH et al. Bolus intrathecal injection of zicono-
R ) to evaluate the option of continuous administration via an implanted
tide (PrialtV
intrathecal drug delivery (ITDD) system: a pilot study. Neuromodulation 2013;16:
576–581; discussion 582.
117. Compton WM, Jones CM, Baldwin GT. Relationship between nonmedical
prescription-opioid use and heroin use. New Engl J Med 2016;174:154–163.
118. Deer TR, Hayek SM et al. The Polyanalgesic Consensus Conference (PACC): recom-
mendations for trialing of intrathecal drug delivery infusion therapy. Neuromodula-
tion 2017;20:133–154.
119. Sultan P, Gutierrez MC, Carvalho B. Neuraxial morphine and respiratory depression:
finding the right balance. Drugs 2011;71:1807–1819.
120. Beatty NC, Arendt KW, Niesen AD et al. Analgesia after cesarean delivery: a retro-
spective comparison of intrathecal hydromorphone and morphine. J Clin Anaesth
2013;25:379–383.
121. Shapiro A, Zohar E, Zaslansky R, Hoppenstein D, Shabat S, Fredman B. The frequen-
cy and timing of respiratory depression in 1524 postoperative patients treated
with systemic or neuraxial morphine. J Clin Anesth 2005;17:537–542.
122. Cowan CM, Kendall JB, Barclay PM et al. Comparison of intrathecal fentanyl and
diamorphine in addition to bupivacaine for Caesarean section under spinal analge-
sia. Br J Anaesth 2002;89:452–458.
123. Duman A, Apiliogullari S, Balasar M, Gurbuz R, Karcioglu M. Comparison of 50
microg and 25 microg doses of intrathecal morphine on postoperative analgesic
requirements in patients undergoing transurethral resection of the prostate with
intrathecal anesthesia. J Clin Anesth 2010;22:329–333.
124. Burton AW, Deer TR, Wallace MS, Rauck RL, Grigsby E. Considerations and method-
ology for trialing ziconotide. Pain Physician 2010;13:23–33.
125. Bolash R, Mekhail M, Multi-center prospective analysis of on-demand intrathecal
morphine bolus dosing among patients with targeted drug delivery systems. Post-
er presented at the North American Neuromodulation Society meeting; Dec. 2015;
Las Vegas, NV.
126. Brogan SE, Winter NB, Okifuji A. Prospective observational study of patient–
128
controlled intrathecal analgesia. Impact on cancer-associated symptoms, break-
through pain control and patient satisfaction. Reg Anes Pain Med 2015;40:369–375.
www.neuromodulationjournal.com C 2017 International Neuromodulation Society
V Neuromodulation 2017; 20: 96–132
127. Fukushima S, Takenami T, Yagishita S, Nara Y, Hoka S, Okamoto H. Neurotoxicity of
intrathecally administered fentanyl in a rat spinal model. Pain Med 2011;12:717–
725.
128. Abut YC, Turkmen AZ, Midi A, Eren B, Yener N, Nurten A. Neurotoxic effects of lev-
obupivacaie and fentanyl on rat spinal cord. Rev Bras Anestesiol 2015;65:27–33.
129. Yaksh TL, Noueihed RY, Durant PAC. Studies of the pharmacology and pathology
of intrathecally administered 4-anilinopiperidine analogues and morphine in rat
and cat. Anesthesiology 1986;64:54–66.
130. Sabbe MB, Grafe MR, Mjanger E, Tiseo PJ, Hill HF, Yaksh TL. Spinal delivery of
sufentanil, alfentanil and morphine in dogs. Physiol Toxicol Investig Anesthesiology
1994;81:899–920.
131. Allen JW, Horais KA, Tozier NA, Yaksh TL. Opiate pharmacology of intrathecal gran-
ulomas. Anesthesiology 2006;105:590–598.
132. Yaksh TL, Steinauer JJ, Veesart SL, Malkmus SA. Alfentanil: correlations between
absence of effect upon subcutaneous mast cells and absence of granuloma forma-
tion after intrathecal infusion in the dog. Neuromodulation 2013;16:459–466; dis-
cussion 466.
133. Rawal N, Nuutinen L, Raj PP et al. Behavioral and histopathologic effects following
intrathecal administration of butorphanol, sufentanil, and nalbuphine in sheep.
Anesthesiology 1991;75:1025–1123.
134. Deer TR, Caraway DL, Kim CK, Dempsey CD, Stewart CD, McNeil KF. Clinical experi-
ence with intrathecal bupivacaine in combination with opioid for the treatment of
chronic pain related to failed back surgery syndrome and metastatic cancer pain
of the spine. Spine J 2002;2:274–278.
135. Hayek SM, Veizi E, Hanes M. Intrathecal hydromorphone and bupivacaine combi-
nation therapy for post-laminectomy syndrome optimized with patient-activated
bolus device. Pain Med 2015; doi http://dx.doi.org/10.1093/pm/pnv021 pnv021.
136. Bernards CM, Hill HF. Physical and chemical properties of drug molecules gov-
erning their diffusion through the spinal meninges. Anesthesiology 1992;77:
750–756.
137. Reina MA, Pulido P, L opez A. Human dural sac [English abstr]. Rev Arg Anestesiol
2007;65:167–184.
138. Reina MA, L opez A, De Andr es JA. Variation of human dura mater thickness
[English abstr]. Rev Esp Anestesiol Reanim 1999;46:344–349.
139. Reina MA, L opez A, Dittmann M, De Andr es JA. Structure of human dura mater
thickness by scanning electron microscopy [English abstr]. Rev Esp Anestesiol
Reanim 1996;43:135–137.
140. Reina MA, Dittmann M, L opez A, van Zundert A. New perspectives in the micro-
scopic structure of human dura mater in the dorso lumbar region. Reg Anesth
1997;22:161–166.
141. Dittmann M, Reina MA, L opez A. Neue ergebnisse bei der darstellung der dura
mater spinalis mittles rasterelektronenmikroskopie [English abstr]. Anaesthesist
1998;47:409–413.
142. Reina MA, L opez A, Dittmann M, De Andr es JA. External and internal surface of
human dura mater by scanning electron microscopy [English abstr]. Rev Esp Anes-
tesiol Reanim 1996;43:130–134.
143. Reina MA, Prats-Galino A, Sola RG, Puigdellıvol-Sanchez A, Arriazu Navarro R, De
Andr es JA. Structure of the arachnoid layer of the human spinal meninges: a barri-
er that regulates dural sac permeability [English abstr]. Rev Esp Anestesiol Reanim
2010;57:486–492.
144. Reina MA, Villanueva MC, L opez A. Human trabecular arachnoids, pia mater and
spinal anesthesia [English abstr]. Rev Arg Anestesiol 2008;66:111–133.
145. Reina MA, L opez A, De Andr es JA. Hypothesis on the anatomical bases of cauda
equine syndrome and transitory radicular irritation syndrome post spinal anesthe-
sia [English abstr]. Rev Esp Anestesiol Reanim 1999;46:99–105.
146. Reina MA, Mach es F, L
opez A, De Andr es JA. The ultrastructure of the spinal arach-
noid in humans and its impact on spinal anesthesia, cauda equina syndrome and
transient neurological syndrome. Tech Reg Anesth Pain Manage 2008;12:153–160.
147. Kershner DE, Binhammer RT. Lumbar Intrathecal ligaments. Clin Anat 2002;15:82–
87.
148. Di Chiro G, Timins EL. Supine myelography and the septum posticum. Radiology
1974;111:319–327.
149. Reina MA, De Leon Casasola O, Villanueva MC, L opez A, Maches F, De Andr es JA.
Ultrastructural findings in human spinal pia mater in relation to subarachnoid
anesthesia. Anesth Analg 2004;98:1479–1485.
150. Reina MA, L opez Garcıa A, de Andr es JA. Anatomical description of a natural perfo-
ration present in the human lumbar pia mater [Spanish]. Rev Esp Anestesiol Reanim
1998;45:4–7.
151. Merchant RE, Low FN. Scanning electron microscopy of the subarachnoid space in
the dog: evidence for a non-hematogeneous origin of the subarachnoid macro-
phages. Am J Anat 1979;156:183–206.
152. Bernards CM. Epidural and intrathecal drug movement. In: Yaksh TN, ed. Spinal
drug delivery. Amsterdam: Elsevier; 1999:239–269.
153. Artru AA, Spinal cerebrospinal fluid chemistry and physiology. In: Yaksh TN, ed. Spi-
nal drug delivery. Amsterdam: Elsevier; 1999:177–237.
154. Quencer RM, Post MJD, Hinks RS. Cine MR in the evaluation of normal and
abnormal CSF flow. Intracranial and intraspinal studies. Neuroradiology 1990;
32:371–391.
155. Wagshul ME, Chen JJ, Egnor MR, McCormack EJ, Roche PE. Amplitude and phase
of cerebrospinal fluid pulsations: experimental studies and review of the literature.
J Neurosurg 2006;104:810–819.
156. Hsu Y, Hettiarachchi HD, Zhu DC, Linninger AA. The frequency and magnitude of
cerebrospinal fluid pulsations influence intrathecal drug distribution: key factors
for interpatient variability. Anesth Analg 2012;115:386–394.

157. Hogan QH, Prost R, Kulier A et al. Magnetic resonance imaging of cerebrospinal
fluid volume and the influence of body habitus and abdominal pressure. Anesthesi-
ology 1996;84:1341–1349.
158. Higuchi H, Hirata J, Adachi Y, Kazama T. Influence of lumbosacral cerebrospinal flu-
id density, velocity, volume and on extent and duration of plain bupivacaine spinal
anesthesia. Anesthesiology 2004;100:106–114.
159. Sullivan JT, Grouper S, Walker MT, Parrish TB, McCarthy RJ, Wong CA. Lumbosacral
cerebrospinal fluid volume in humans using three-dimensional magnetic reso-
nance imaging. Anesth Analg 2006;103:1306–1310.
160. Edsbagge M, Starck G, Zetterberg H, Ziegelitz D, Wikkelso C. Spinal CSF volume in
healthy elderly individuals. Clin Anat 2011;24:733–740.
161. Puigdellıvol-Sanchez A, Prats-Galino A, Reina MA et al. Tridimensional magnetic
resonance image of structures enclosed in the spinal canal relevant to anesthetists
and estimation of the lumbosacral CSF volume. Acta Anaesth Belg 2011;62:37–45.
162. Prats-Galino A, Reina MA, Puigdellıvol-Sanchez A, Juanes M endez JA, De Andr es
JA, Collier CB. Cerebrospinal fluid volume and nerve root vulnerability during lum-
bar puncture or spinal anaesthesia at different vertebral levels. Anaesth Intensive
Care 2012;40:643–647.
163. Prats-Galino A, Reina MA, Mavar Haramija M, Puigdellivol-Sanchez A, Juanes
Mendez JA, De Andres JA. 3D interactive model of lumbar spinal structures of
anesthetic interest. Clin Anat 2015;28:205–212.
164. Papisov MI, Belov VV, Gannon KS. Physiology of the intrathecal bolus: the leptome-
ningeal route for macromolecule and particle delivery to CNS. Mol Pharm 2013;10:
1522–1532.
165. Dreha-Kulaczewski S, Joseph AA, Merboldt KD, Ludwig HC, Gartner J, Frahm J.
Inspiration is the major regulator of human CSF flow. J Neurosci 2015;35:2485–
2491.
166. Brinker T, Stopa E, Morrison J, Klinge P. A new look at cerebrospinal fluid circula-
tion. Fluids Barriers CNS 2014;11:10.
167. Kuttler A, Dimke T, Kern S, Helmlinger G et al. Understanding pharmacokinetics
using realistic computational models of fluid dynamics: biosimulation of drug dis-
tribution within the CSF space for intrathecal drugs. J Pharmacokinet Pharmacodyn
2010;37:629–664.
168. Nelissen RM. Fluid mechanics of intrathecal drug delivery [thesis]. Lausanne: Federal
Institute of Technology, 2008.
169. Henry-Feugeas MC, Idy-Peretti L, Blanchet B et al. Temporal and spatial assessment
of normal cerebrospinal fluid dynamics with MR imaging. Magn Reson Imaging
1993;11:1107–1118.
170. Wallace M, Yaksh TL. Characteristics of distribution of morphine and metabolites
in cerebrospinal fluid and plasma with chronic intrathecal morphine infusion in
humans. Anesth Analg 2012;114:797–804.
171. Buchser E, Durrer A, Chedel E et al. Efficacy of intrathecal bupivacaine: how impor-
tant is the flow rate? Pain Med 2004;5:248–252.
172. Mironer E, Haasis J, Chapple I et al. Efficacy and safety of intrathecal opioid/bupiva-
caine mixture in chronic nonmalignant pain: a double blind, randomized, cross-
over, multicenter study by the National Forum of Independent Pain Clinicians
(NFIPC). Neuromodulation 2002;5:208–213.
173. Bernards CM, Shen DD, Sterling ES et al. Epidural, cerebrospinal fluid, and plasma
pharmacokinetics of epidural opioids (part 1): differences among opioids. Anesthe-
siology 2003;99:455–465.
174. Dahm PO, Nitescu PV, Appelgren LK, Curelaru ID. Intrathecal infusion of bupiva-
caine with or without buprenorphine relieved intractable pain in three patients
with vertebral compression fractures caused by osteoporosis. Reg Anest Pain Med
1999;24:352–357.
175. Wall PD, Shortland P. Long-range afferents in the rat spinal cord. 1. Numbers,
distances and conduction velocities. Phil Trans R Soc London B Biol Sci 1991;
334:85–93.
176. Shortland P, Wall PD. Long-range afferents in the rat spinal cord. II. Arborizations
that penetrate grey matter. Phil Trans R Soc London B Biol Sci 1992;337:445–455.
177. Dahm P, Nitescu P, Appelgren L, Curelaru I. Efficacy and technical complications of
long-term continuous intraspinal infusions of opioid and/or bupivacaine in refrac-
tory nonmalignant pain: a comparison between the epidural and the intrathecal
approach with externalized or implanted catheters and infusion pumps. Clin J Pain
1998;14:4–16.
178. Yaksh TL, Woller SA, Ramachandran R, Sorkin LS. The search for novel analgesics:
targets and mechanisms. F1000 Prime Rep 2015;26:7–56.
179. Kroin JS. Intrathecal drug administration. Present use and future trends. Clin Phar-
macokin 1992;22:319–326.
180. Rennels ML, Blaumanis OR, Grady PA. Rapid solute transport throughout the brain
via paravascular fluid pathways. Adv Neurol 1990;52:431–439.
181. Bernards CM. Understanding the physiology and pharmacology of epidural and
intrathecal opioids. Best Pract Res Clin Anaesthesiol 2002;16:489–505.
182. Nordberg G. Pharmacokinetic aspects of spinal morphine analgesia. Acta Anaesth
Scand (Suppl) 1984;79:1–38.
183. Gudeman J, Jozwiakowski M, Chollet J, Randell M. Potential risks of pharmacy
compounding. Drugs R D 2013;13:1–8.
184. Ghafoor VL, Epshteyn M, Carlson GH, Terhaar DM et al. Intrathecal drug therapy
for long-term patient management. Am J Heath Syst Pharm 2007;64:2447–2461.
185. Rezai A, Kloth D, Hansen H, Schultz D, Thompson S, Deer TR. Physician response to
medtronic’s position on the use of off-label medications in the synchromed pump.
Pain Phys 2013;16:415–417.
186. Medtronic product bulletin: summary of approved drugs Medtronic’s SynchroMed II
Infusion System. Minneapolis, MN: Medtronic plc, 2012.
187. Kim P, Case series of distal catheter obstruction. Presented at: NANS Annual Meet-
ing; 2011; Las Vegas, NV.
www.neuromodulationjournal.com C 2017 International Neuromodulation Society
V Neuromodulation 2017; 20: 96–132
INTRATHECAL THERAPY BEST PRACTICES AND GUIDELINES
188. Medtronic. Intrathecal distal end catheter occlusions as a result of pH & salt con-
centration gradients between delivery solution & cerebrospinal fluid (CSF). Minne-
apolis, MN: Medtronic plc; 2009.
189. Inturrisi CE. Clinical pharmacology of opioids for pain. Clin J Pain 2002;18(4 Suppl):
S3–13.
190. Yaksh TL, Allen JW, Veesart SL et al. Role of meningeal mast cells in intrathe-
cal morphine-evoked granuloma formation. Anesthesiology 2013;118:
664–678.
191. Yaksh TL, Horais KA, Tozier NA et al. Chronically infused intrathecal morphine in
dogs. Anesthesiology 2003;99:174–187.
192. Michael A, Buffen E, Rauck R, Anderson W, McGirt M, Mendenhall HV. An in vivo
canine study to assess granulomatous responses in the MedStream Programmable
Infusion System (TM) and the SynchroMed II Infusion SystemV R . Pain Med 2012;13:
175–184.
193. Gradert TL, Baze WB, Satterfield WC, Hildebrand KR, Johansen MJ, Hassenbusch SJ.
Safety of chronic intrathecal morphine infusion in a sheep model. Anesthesiology
2003;99:188–198.
194. Yaksh TL. Spinal delivery and assessment of drug safety. In: Bolon B, Butt MT, eds.
Fundamental neuropathology for pathologists and toxicologists, Chapter 27. Philadel-
phia: John Wiley and Sons; 2011:451–462.
195. Atli A, Theodore BR, Turk DC, Loeser JD. Intrathecal opioid therapy for chronic non-
malignant pain: a retrospective cohort study with 3-year follow-up. Pain Med 2010;
11:1010–1016.
196. Berg A, Barsa J, Deer T et al. Efficacy of morphine sulfate infusion via the Prom-
etraVR Intrathecal Infusion Pump. A prospective multicenter evaluation: PB238
[abstract]. Pain Pract 2009;9(Suppl.1):159.
197. Kongkam P, Wagner DL, Sherman S et al. Intrathecal narcotic infusion pumps for
intractable pain of chronic pancreatitis: a pilot series. Am J Gastroenterol 2009;104:
1249–1255.
198. Shaladi A, Saltari MR, Piva B et al. Continuous intrathecal morphine infusion in
patients with vertebral fractures due to osteoporosis. Clin J Pain 2007;23:511–
517.
199. Saltari MR, Shaladi A, Piva B et al. The management of pain from collapse of osteo-
porotic vertebrae with continuous intrathecal morphine infusion. Neuromodulation
2007;10:167–176.
200. Reig E, Abejon D. Continuous morphine infusion: a retrospective study of efficacy,
safety, and demographic variables. Neuromodulation 2009;12:122–129.
201. Raphael JH, Duarte RV, Southall JL, Nightingale P, Kitas GE. Randomised, double-
blind controlled trial by dose reduction of implanted intrathecal morphine delivery
in chronic non-cancer pain. BMJ Open 2013;3:e003061.
202. Kumar P, Sunkaraneni S, Sirohi S, Dighe SV, Walker EA, Yoburn BC. Hydromor-
phone efficacy and treatment protocol impact on tolerance and mu-opioid recep-
tor regulation. Eur J Pharmacol 2008;597:39–45.
203. Johansen MJ, Satterfield WC, Baze WB, Hildebrand KR, Gradert TL, Hassenbusch SJ.
Continuous intrathecal infusion of hydromorphone: safety in the sheep model and
clinical implications. Pain Med 2004;5:14–25.
204. Hoederath P, Gautschi OP, Land M, Hildebrandt G, Fournier JY. Formation of two
consecutive intrathecal catheter tip granulomas within nine months. Cen Eur Neu-
rosurg 2010;71:39–42.
205. Ramsey CN, Owen RD, Witt WO, Grider JS. Intrathecal granuloma in a patient
receiving high dose hydromorphone. Pain Physician 2008;11:369–373.
206. Ruan X, Liu H, Couch JP, Wang F, Chiravuri S. Recurrent cellulitis associated with
long-term intrathecal opioid infusion therapy: a case report and review of the liter-
ature. Pain Med 2010;11:972–976.
207. Sirohi S, Dighe SV, Walker EA, Yoburn BC. The analgesic efficacy of fentanyl: rela-
tionship to tolerance and mu-opioid receptor regulation. Pharmacol Biochem
Behav 2008;91:115–120.
208. Medel R, Pouratian N, Elias WJ. Catheter-tip mass mimicking a spinal epidural
hematoma. J Neurosurg Spine 2010;12:66–71.
209. Colpaert FC, Leysen JE, Michiels M, van den Hoogen RH. Epidural and intravenous
sufentanil in the rat: analgesia, opiate receptor binding, and drug concentrations
in plasma and brain. Anesthesiology 1986;65:41–49.
210. Gupta A, Martindale T, Christo PJ. Intrathecal catheter granuloma associated with
continuous sufentanil infusion. Pain Med 2010;11:847–852.
211. Reich DL, Silvay G. Ketamine: an update on the first twenty-five years of clinical
experience. Can J Anaesth 1989;36:186–197.
212. Yaksh TL, Tozier N, Horais KA et al. Toxicology profile of N-methyl-D-aspartate
antagonists delivered by intrathecal infusion in the canine model. Anesthesiology
2008;108:938–949.
213. Walker SM, Westin BD, Deumens R, Grafe M, Yaksh TL. Effects of intrathecal keta-
mine in the neonatal rat: evaluation of apoptosis and long-term functional out-
come. Anesthesiology 2010;113:147–159.
214. Wilson JA, Nimmo AF, Fleetwood-Walker SM, Colvin LA. A randomised double
blind trial of the effect of pre-emptive epidural ketamine on persistent pain after
lower limb amputation. Pain 2008;135:108–118.
215. Sato C, Okabe T, Nakanishi K, Sakamoto A. A case of cancer pain management by
long-term intrathecal PCA. J Nippon Med School 2010;77:333–337.
216. Karpinski N, Dunn J, Hansen L, Masliah E. Subpial vacuolar myelopathy after intra-
thecal ketamine: report of a case. Pain 1997;73:103–105.
217. Rizzo CC, Luretti GR, Mattos AL. The combination of methadone and dexametha-
sone for cancer pain. Eur J Pain 2009;13(Suppl. 1):S258.
218. Mironer YE, Tollison CD. Methadone in the intrathecal treatment of chronic non-
129
malignant pain resistant to other neuroaxial agents: the first experience. Neuromo-
dulation 2001;4:25–31.
 DEER ET AL.
219. Zamponi GW, Striessnig J, Koschak A, Dolphin AC. The physiology, pathology, and
pharmacology of voltage-gated calcium channels and their future therapeutic
potential. Pharmacol Rev 2015;67:821–870.
220. Takasusuki T, Yaksh TL. Regulation of spinal substance p release by intrathecal cal-
cium channel blockade. Anesthesiology 2011;115:153–164.
221. Skov MJ, Beck JC, de Kater AW, Shopp GM. Nonclinical safety of ziconotide:
an intrathecal analgesic of a new pharmaceutical class. Int J Toxicol 2007;26:
411–421.
222. Rainov NG, Heidecke V, Burkert W. Long-term intrathecal infusion of drug combi-
nations for chronic back and leg pain. J Pain Symptom Manage 2001;22:862–871.
223. Butterworth JF, Strichartz GR. Molecular mechanisms of local anesthesia: a review.
Anesthesiology 1990;72:711–734.
224. Covino BG. Pharmacology of local anaesthetic agents. Br J Anaesth 1986;58:701–716.
225. Scott DB, McClure JH, Giasi RM, Seo J, Covino BG. Effects of concentration of local
anaesthetic drugs in extradural block. Br J Anaesth 1980;52:1033–1037.
226. Cohen SE, Yeh JY, Riley ET, Vogel TM. Walking with labor epidural analgesia: the
impact of bupivacaine concentration and a lidocaine-epinephrine test dose. Anes-
thesiology 2000;92:387–392.
227. Abrahams M, Higgins P, Whyte P, Breen P, Muttu S, Gardiner J. Intact propriocep-
tion and control of labour pain during epidural analgesia. Acta Anaesthesiol Scand
1999;43:46–50.
228. Dahm P, Nitescu P, Appelgren L, Curelaru I. High thoracic/low cervical, long-term intra-
thecal (i.t.) infusion of bupivacaine alleviates “refractory” pain in patients with unstable
angina pectoris. Report of 2 cases. Acta Anaesthesiol Scand 1998;42:1010–1017.
229. Dahm PO, Nitescu PV, Appelgren LK, Curelaru II. Long-term intrathecal (i.t.) infu-
sion of bupivacaine relieved intractable pain and spasticity in a patient with multi-
ple sclerosis. Eur J Pain 1998;2:81–85.
230. Lundborg C, Dahm P, Nitescu P, Appelgren L, Curelaru I. Clinical experience using
intrathecal (IT) bupivacaine infusion in three patients with complex regional pain
syndrome type I (CRPS-I). Acta Anaesthesiol Scand 1999;43:667–678.
231. Akerman B, Arwestr€ om E, Post C. Local anesthetics potentiate spinal morphine
antinociception. Anesth Analg 1988;67:943–948.
232. Li DF, Bahar M, Cole G, Rosen M. Neurological toxicity of the subarachnoid infusion
of bupivacaine, lignocaine or 2-chloroprocaine in the rat. BrJ Anaesth 1985;57:
424–429.
233. Kroin JS, McCarthy RJ, Penn RD, Kerns JM, Ivankovich AD. The effect of chronic
subarachnoid bupivacaine infusion in dogs. Anesthesiology 1987;66:737–742.
234. Wagemans MF, van der Valk P, Spoelder EM, Zuurmond WW, de Lange JJ. Neuro-
histopathological findings after continuous intrathecal administration of morphine
or a morphine/bupivacaine mixture in cancer pain patients. Acta Anaesthesiol
Scand 1997;41:1033–1038.
235. Sjoberg M, Karlsson PA, Nordborg C et al. Neuropathologic findings after long-
term intrathecal infusion of morphine and bupivacaine for pain treatment in can-
cer patients. Anesthesiology 1992;76:173–186.
236. Ortner CM, Posch M, Roessler B et al. On the ropivacaine-reducing effect of low-
dose sufentanil in intrathecal labor analgesia. Acta Anaesthesiol Scand 2010;54:
1000–1006.
237. Tejwani GA, Rattan AK, McDonald JS. Role of spinal opioid receptors in the antino-
ciceptive interactions between intrathecal morphine and bupivacaine. Anesth
Analg 1992;74:726–734.
238. Parpaglioni R, Baldassini B, Barbati G, Celleno D. Adding sufentanil to levobupiva-
caine or ropivacaine intrathecal anaesthesia affects the minimum local anaesthetic
dose required. Acta Anaesthesiol Scand 2009;53:1214–1220.
239. Levin A, Datta S, Camann WR. Intrathecal ropivacaine for labor analgesia: a com-
parison with bupivacaine. Anesth Analg 1998;87:624–627.
240. van Dongen RT, Crul BJ, van Egmond J. Intrathecal coadministration of bupiva-
caine diminishes morphine dose progression during long-term intrathecal infusion
in cancer patients. Clin J Pain 1999;15:166–172.
241. Ilias W, le Polain B, Buchser E, Demartini L. Patient-controlled analgesia in chronic
pain patients: experience with a new device designed to be used with implanted
programmable pumps. Pain Pract 2008;8:164–170.
242. Lundborg C, Dahm P, Nitescu P, Biber B. High intrathecal bupivacaine for severe
pain in the head and neck. Acta Anaesthesiol Scand 2009;53:908–913.
243. Yaksh TL. Pharmacology of spinal adrenergic systems which modulate spinal noci-
ceptive processing. Pharmacol Biochem Behav 1985;22:845–858.
244. Guevara-L opez U, Aldrete JA, Covarrubias-G omez A, Hernandez-Pando RE, L opez-
Mu~ noz FJ. Absence of histological changes after the administration of a continu-
ous intrathecal clonidine in Wistar rats. Pain Pract 2009;9:122–129.
245. Yaksh TL, Rathbun M, Jage J, Mirzai TH, Grafe M, Hiles RA. Pharmacology and toxi-
cology of chronically infused epidural clonidine HCl in dogs. Fundam Appl Toxicol
1994;23:319–335.
246. Bevacqua BK, Fattouh M, Backonja M. Depression, night terrors, and insomnia
associated with long-term intrathecal clonidine therapy. Pain Pract 2007;7:36–38.
247. Rauck RL, North J, Eisenach JC. Intrathecal clonidine and adenosine: effects on
pain and sensory processing in patients with chronic regional pain syndrome. Pain
2015;156:88–95.
248. Engelman E, Marsala C. Efficacy of adding clonidine to intrathecal morphine in
acute postoperative pain: meta-analysis. Br J Anesth 2013;110:21–27.
249. Koman G, Alfieri A, Rachinger J et al. Erectile dysfunction as rare side effect in the
simultaneous intrathecal application of morphine and clonidine. Pain Physician
2012;15:E523–E526.
250. Dupoiron D, Richard H, Chabert-Desnot V et al. In vitro stability of low-
130
concentration ziconotide alone or in admixtures in intrathecal pumps. Neuromodu-
lation 2014;17:472–482.
www.neuromodulationjournal.com C 2017 International Neuromodulation Society
V Neuromodulation 2017; 20: 96–132
251. Munro G, Ahring PK, Mirza NR. Developing analgesics by enhancing spinal inhibi-
tion after injury: GABA-A receptor subtypes as novel targets. Trends Pharmacol Sci
2009;30:453–459.
252. Zeilhofer HU, Witschi R, H€ osl K. Subtype-selective GABAA receptor mimetics—
novel antihyperalgesic agents? J Mol Med 2009;87:465–469.
253. Chiari A, Yaksh TL, Myers RR et al. Preclinical toxicity screening of intrathecal aden-
osine in rats and dogs. Anesthesiology 1999;91:824–832.
254. Sabbe MB, Grafe MR, Pfeifer BL, Mirzai TH, Yaksh TL. Toxicology of baclofen contin-
uously infused into the spinal intrathecal space of the dog. Neurotoxicology 1993;
14:397–410.
255. Koulousakis A, Kuchta J. Intrathecal antispastic drug application with implantable
pumps: results of a 10 year follow-up study. Acta Neurochir Suppl 2007;97 (Pt. 1):
181–184.
256. Deer TR, Raso LJ, Garten TG. Inflammatory mass of an intrathecal catheter in
patients receiving baclofen as a sole agent: a report of two cases and a
review of the identification and treatment of the complication. Pain Med 2007;
8:259–262.
257. Deer TR, Raso LJ, Coffey RJ, Allen JW. Intrathecal baclofen and catheter tip inflam-
matory mass lesions (granulomas): a reevaluation of case reports and imaging
findings in light of experimental, clinicopathological, and radiological evidence.
Pain Med 2008;9:391–395.
258. van der Plas AA, Marinus J, Eldabe S, Buchser E, van Hilten JJ. The lack of effi-
cacy of different infusion rates of intrathecal baclofen in complex regional
pain syndrome: a randomized, double-blind, crossover study. Pain Med 2011;
12:459–465.
259. Saulino M, Burton AW, Danyo DA, Frost S, Glanzer J, Solanki DR. Intrathecal zicono-
tide and baclofen provide pain relief in seven patients with neuropathic pain and
spasticity: case reports. Eur J Phys Rehabil Med 2009;45:61–67.
260. Lind G, Schechtmann G, Winter J, Meyerson BA, Linderoth B. Baclofen-enhanced
spinal cord stimulation and intrathecal baclofen alone for neuropathic pain: long-
term outcome of a pilot study. Eur J Pain 2008;12:132–136.
261. Heetla HW, Staal MJ, Kliphuis C, van Laar T. The incidence and management of tol-
erance in intrathecal baclofen therapy. Spinal Cord 2009;47:751–756.
262. Dones I, Broggi G. A case of very long-term appearing drug tolerance to intrathe-
cal baclofen. J Neurosurg Sci 2010;54:77–78.
263. Hansen CR, Gooch JL, Such-Neibar T. Prolonged, severe intrathecal baclofen
withdrawal syndrome: a case report. Arch Phys Med Rehabil 2007;88:
1468–1471.
264. Salazar ML, Eiland LS. Intrathecal baclofen withdrawal resembling serotonin
syndrome in an adolescent boy with cerebral palsy. Pediatr Emerg Care 2008;
24:691–693.
265. D’Aleo G, Cammaroto S, Rifici C et al. Hallucinations after abrupt withdrawal of
oral and intrathecal baclofen. Funct Neurol 2007;22:81–88.
266. Ross JC, Cook AM, Stewart GL, Fahy BG. Acute intrathecal baclofen withdrawal: a
brief review of treatment options. Neurocrit Care 2011;14:103–108.
267. Fernandes P, Dolan L, Weinstein SL. Intrathecal baclofen withdrawal syndrome fol-
lowing posterior spinal fusion for neuromuscular scoliosis: a case report. Iowa
Orthop J 2008;28:77–80.
268. Duhon BS, MacDonald JD. Infusion of intrathecal baclofen for acute withdrawal.
Technical note. J Neurosurg 2007;107:878–880.
269. Dalton C, Keenan E, Stevenson V. A novel cause of intrathecal baclofen overdos-
age: lessons to be learnt. Clin Rehabil 2008;22:188–190.
270. Mercadante S, Intravaia G, Villari P et al. Intrathecal treatment in cancer
patients unresponsive to multiple trials of systemic opioids. Clin J Pain 2007;
23:793–798.
271. Krames E. A history of intraspinal analgesia, a small and personal journey. Neuro-
modulation 2012;15:172–193.
272. Raffaeli W, Righetti D, Caminiti A et al. Implantable intrathecal pumps for the treat-
ment of noncancer chronic pain in the elderly population: drug dose and clinical
efficacy. Neuromodulation 2008;11:33–39.
273. Webster LR, Fakata KL, Charapata S, Fisher R, MineHart M. Open-label, multi-
center study of combined intrathecal morphine and ziconotide: addition of
morphine in patients receiving ziconotide for severe chronic pain. Pain Med
2008;9:282–290.
274. Shields DC, Palma C, Khoo LT, Ferrante FM. Extramedullary intrathecal catheter
granuloma adherent to the conus medullaris presenting as cauda equina syn-
drome. Anesthesiology 2005;102:1059–1061.
275. PrometraV R II programmable pump. Mt. Olive, NJ: Flowonix Medical Inc., 2012.
276. MRI Guidelines. http://professional.medtronic.com/pt/neuro/idd/ind/mri-guide-
lines/#.Vq5SgjYrIUE.
277. Rauck R, Deer T, Rosen S et al. Accuracy and efficacy of intrathecal administration
of morphine sulfate for treatment of intractable pain using the Prometra program-
mable pump. Neuromodulation 2010;13:102–108.
278. Rauck R, Deer T, Rosen S. Long-term follow-up of a novel implantable programma-
ble infusion pump. Neuromodulation 2013;16:163–167.
279. Rosen SM, Bromberg TA, Padda G et al. Intrathecal administration of InfumorphV R
vs compounded morphine for treatment of intractable pain using the PrometraV R
programmable pump. Pain Med 2013;14:865–873.
280. Medtronic. Product Advisories. http://professional.medtronic.com/pt/neuro/idd/
ind/product-advisories/index.htm?cmpid5URL_Neuro_HCP_iddadvisories_-#.
Vet6E7eBIZE.
281. Tagen KM, Hsu Y, Zhu DC, Linninger AA. CNS wide simulation of low resistance
and drug transport due to spinal microanatomy. J Biomech 2015; http://dx.doiorg/
10.1016/Jbiomech.205.02.018.

282. Perruchoud C, Eldabe S, Durrer A et al. Effects of flow rate modifications on
reported analgesia and quality of life in chronic pain patients treated with continu-
ous intrathecal drug therapy. Pain Med 2011;12:571–576.
283. Heittiarachchi HDM, Hsu Y, Harris TJ, Linninger AA. The effect of pulsatile
flow on intrathecal drug delivery in the spinal canal. J Biomed Eng 2011;39:
2592–2602.
284. Anderson VC, Burcheil KJ, Cooke B. A prospective randomized trial of intrathecal
injection vs epidural infusion in the selection of patients for continuous intrathecal
opioid therapy. Neuromodulation 2003;6:142–152.
285. Grider JS, Harned ME, Sloan PA. Patient selection and trialing techniques utilizing
low-dose intrathecal morphine for chronic nonmalignant pain: a report of two
cases. J Opioid Manag 2010;6:371–376.
286. Reck T, Change EC, Bechir M et al. Applying a part of the daily dose as boli may
improve intrathecal opiod therapy in patients with chronic pain. Neuoromodulation
2016; doi: 10.1111/ner.12391.
287. Nelson DV, Kennington M, Novy DM, Squitieri P. Psychological selection criteria for
implantable spinal cord stimulators. Pain Forum 1996;5:93–103.
288. Saulino M, Kim PS, Shaw E. Practical considerations and patient selection for intra-
thecal drug delivery in the management of chronic pain. J Pain Res 2014;7:627–
638.
289. Schatman ME. Selection criteria for intrathecal opioid therapy: a re-examination of
the “science.” Pract Pain Manag 2005; http://www.practicalpainmanagement.com/
treatments/interventional/pumps/selection-criteria-intrathecal-opioid-therapy-
re-examination-science.
COMMENTS
This refreshing update to the Polyanalgesic Consensus Conference
statement assembles current evidence in basic science, anatomy, physi-
ology, engineering and patient care surrounding targeted drug delivery.
Thirty concise and tangible recommendations are accompanied by use-
ful flow charts, reference tables and decision trees providing guidance
which can be considered for contemporary pain care in patients with
targeted drug delivery systems.
The intrathecal pharmacopeia has expanded in this iteration with divi-
sion of first line agents into 1A and 1B categories, and drug choices are
now categorized by localized or diffuse pain conditions, as well as cancer
or non-cancer related diagnosis. Most notably, the current version largely
combines the previous neuropathic/nociceptive divisions presented in
the 2012 consensus conference (1), instead focusing on anatomic loca-
tion: diffuse or localized pain in patients with noncancer related pain.
Some guidance is sure to prove controversial; while, at times, other
recommendations seem anecdotal. Ziconotide maintains a first line
position in both algorithms, yet sales of ziconotide demonstrate a
shrinking year-over-year minority of real-world use (2), and may continue
to decrease with increasing attention on healthcare cost containment.
Sufentanil monotherapy is recommended for cancer and other terminal
pain conditions, but there is never a recommendation to use fentanyl
monotherapy in these same patients. Bupivacaine monotherapy is
placed second line for localized non-cancer pain, but never appears in
isolation for the treatment of cancer related pain conditions. At times,
the reader is left wondering: what is anecdote, experience or evidence?
Amongst the most demanding components of physicianship is the
task of applying a paucity of high-quality clinical evidence to each unique
patient at the bedside. Regardless of the comprehensiveness of any con-
sensus committee recommendation, clinicians will continue to be
required to formulate treatment plans in the absence of a randomized
double-blind placebo controlled trial mimicking the patient before them.
Perhaps the most notable conclusion upon reading the consensus state-
ment is how few rigorous, well designed clinical studies exist to support
clinical practice. With a growing number of patients suffering with chron-
ic pain (3), and an observed quadrupling of deaths associated with oral
opioids (4), the time is ripe for interventional pain physicians to demand
funding of high-quality studies to support contemporary pain care.
Robert Bolash, MD
Cleveland, OH, USA
www.neuromodulationjournal.com C 2017 International Neuromodulation Society
V Neuromodulation 2017; 20: 96–132
INTRATHECAL THERAPY BEST PRACTICES AND GUIDELINES
REFERENCES
1. Deer TR, Prager J, Levy R, Rathmell J, Buchser E, Burton A et al. Polya-
nalgesic Consensus Conference 2012: recommendations for the
management of pain by intrathecal (intraspinal) drug delivery: report
of an interdisciplinary expert panel. Neuromodulation. 2012 Sep-
Oct;15:436–64.
2. Jazz Pharmaceuticals (2015) Jazz Pharmaceuticals Announces Full
Year And Fourth Quarter 2014 Financial Results [Press Release]
Retrieved from: http://www.prnewswire.com/news-releases/jazz-
pharmaceuticals-announces-full-year-and-fourth-quarter-2014-financial-
results-300040639.html
3. Institute of Medicine (US) Committee on Advancing Pain Research,
Care, and Education. Relieving Pain in America: A Blueprint for Trans-
forming Prevention, Care, Education, and Research. Washington (DC):
National Academies Press (US); 2011. 2, Pain as a Public Health Chal-
lenge. Available from: http://www.ncbi.nlm.nih.gov/books/NBK92516/
4. CDC. Wide-ranging online data for epidemiologic research (WONDER).
Atlanta, GA: CDC, National Center for Health Statistics; 2016. Available
at http://wonder.cdc.gov
***
For many years the Polyanalgesic Consensus Conference recommen-
dations are an important contribution to the management of intrathecal
therapy. Again, the new version has been thoroughly revised and
includes all relevant features of intrathecal therapy. Of course some
points have to remain unanswered and one issue that can always be
argued on is the "lines" of intrathecal drugs to choose, especially since
there is an almost infinite number of possible combinations. In the end,
every physician has to make an individual choice based on pain- and
patient characteristics, and a recommendation can only be somewhat
of a blueprint.
Tim Reck, MD
Nottwil, Switzerland
***
The Polyanalgesic Consensus Conference recommendations on intra-
thecal drug delivery is a living document that continues to impact clini-
cal practice and guide future research. The fifth iteration of these
recommendations are patient-centered and crafted in a manner that
condenses the last four years of IDDS research into a practical resource
for practicing clinicians.
Critical assessment of evidence from multiple studies are necessary
when making decisions in day to day practice. Medical literature is
immense and not all information is clinically useful. The 2016 guidelines
systematically ranked evidence based on the hierarchy of studies. Expert
consensus was measured and the degree of recommendations were
defined. Research gaps were also identified which will hopefully lead to
scholarly queries and future investigations. I commend the authors for
their work.
While appropriate patient selection has been identified as one of the
most important factors for a successful outcome when initiating intra-
thecal therapy, long term management of the intrathecal device poses
several challenges. Pain patterns may change due to an ongoing disease
process or changes in psychological status years later. Non-medical
determinants of health, such as social and economic status, may strain
the delivery of needed services to maintain an intrathecal device.
Guidance regarding long term management of the intrathecal device
was briefly addressed in this manuscript, but more detailed recommen-
131
dations would be useful. For example, at what point should a clinician
 DEER ET AL.

consider discontinuation of intrathecal therapy because of lack or loss of
efficacy? Is there a consensus on the timing of intrathecal therapy rota-
tion from one opioid to another or from an opioid to Ziconotide or to
combination therapy? Is there consensus on the benefit or lack there of
for trailing a different intrathecal medication before rotation? Many of us
have been faced with these decisions while in the trenches of clinical
practice. We want to salvage this therapy which was at one point effica-
cious, but never ask ourselves when is enough really enough? For many
clinicians with large intrathecal pump practices, long term management
132
is probably even more complex than the initial identification of an
appropriate candidate. Successful outcomes rely upon discovering the
science behind the art of managing the complexities of a long term
intrathecal pump.
Brian Bruel, MD, MBA
Houston, TX, USA
Comments not included in the Early View version of this paper.

www.neuromodulationjournal.com C 2017 International Neuromodulation Society

V Neuromodulation 2017; 20: 96–132